Hi All,

I think I found somewhat of a showstopper bug.  I am testing using the
latest Csound code from CVS and in rendernig this in realtime it cuts
off the last bit of the last note with a pop.  If I render this to
disk it is fine.  I'm on Windows XP.  If I use MME, it cuts off fairly
early in the last note; if I use PortAudio, it cuts off near the end
of the note.

The score renders to completion with no errors reported. My guess is
that Csound isn't waiting for the audio to flush before exiting, but I
am just guessing.  Does anyone have any ideas?  Could someone try this
on other platforms to see if the problem occurs there too?

Thanks,
steven

Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using Alsa and
Portaudio

But my ears are not good
==John ff

> Hi All,
>
> I think I found somewhat of a showstopper bug.  I am testing using the
> latest Csound code from CVS and in rendernig this in realtime it cuts
> off the last bit of the last note with a pop.  If I render this to
> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts off fairly
> early in the last note; if I use PortAudio, it cuts off near the end
> of the note.
>
> The score renders to completion with no errors reported. My guess is
> that Csound isn't waiting for the audio to flush before exiting, but I
> am just guessing.  Does anyone have any ideas?  Could someone try this
> on other platforms to see if the problem occurs there too?
>
> Thanks,
> steven
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for
> Conference
> attendees to learn about information security's most important issues
> through
> interactions with peers, luminaries and emerging and established
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>



------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

  Same here OSX with doubles version (I have not yet built a floats  
one) (I hear no problems). I'll build the floats and test.

Victor
On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:

> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using Alsa and
> Portaudio
>
> But my ears are not good
> ==John ff
>
>> Hi All,
>>
>> I think I found somewhat of a showstopper bug.  I am testing using  
>> the
>> latest Csound code from CVS and in rendernig this in realtime it cuts
>> off the last bit of the last note with a pop.  If I render this to
>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts off  
>> fairly
>> early in the last note; if I use PortAudio, it cuts off near the end
>> of the note.
>>
>> The score renders to completion with no errors reported. My guess is
>> that Csound isn't waiting for the audio to flush before exiting,  
>> but I
>> am just guessing.  Does anyone have any ideas?  Could someone try  
>> this
>> on other platforms to see if the problem occurs there too?
>>
>> Thanks,
>> steven
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently  
>> attracts the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important issues
>> through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Same with floats version, no audible problem, disk and dac versions  
are the same.
On 19 Jan 2010, at 08:45, Victor Lazzarini wrote:

> Same here OSX with doubles version (I have not yet built a floats
> one) (I hear no problems). I'll build the floats and test.
>
> Victor
> On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:
>
>> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using Alsa and
>> Portaudio
>>
>> But my ears are not good
>> ==John ff
>>
>>> Hi All,
>>>
>>> I think I found somewhat of a showstopper bug.  I am testing using
>>> the
>>> latest Csound code from CVS and in rendernig this in realtime it  
>>> cuts
>>> off the last bit of the last note with a pop.  If I render this to
>>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts off
>>> fairly
>>> early in the last note; if I use PortAudio, it cuts off near the end
>>> of the note.
>>>
>>> The score renders to completion with no errors reported. My guess is
>>> that Csound isn't waiting for the audio to flush before exiting,
>>> but I
>>> am just guessing.  Does anyone have any ideas?  Could someone try
>>> this
>>> on other platforms to see if the problem occurs there too?
>>>
>>> Thanks,
>>> steven
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently
>>> attracts the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important  
>>> issues
>>> through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts
>> the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important
>> issues through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hi,

I can confirm this problem on Linux. I ran through QuteCsound but
using the terminal, to make sure it was not an problem with
QuteCsound. I used Jack to connect the output of Csound to Audacity,
there were no xruns and I used this csd:






sr = 44100
ksmps = 128
nchnls = 2
0dbfs = 1

instr 2
aamp linen 0.2, 0.02, p3, 0.02
aout oscil aamp, 440, 1
outs aout, aout
endin



f 1 0 1024 10 1
i 2 0 2
i 2 3 2
e



The tails of the two notes are shown in the images. It would seem that
about 5ms were cut off. I was using -b1024 -B2048 and jack buffer size
of 2048.

Cheers,
Andrés


On Tue, Jan 19, 2010 at 8:58 AM, Victor Lazzarini
 wrote:
> Same with floats version, no audible problem, disk and dac versions
> are the same.
> On 19 Jan 2010, at 08:45, Victor Lazzarini wrote:
>
>> Same here OSX with doubles version (I have not yet built a floats
>> one) (I hear no problems). I'll build the floats and test.
>>
>> Victor
>> On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:
>>
>>> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using Alsa and
>>> Portaudio
>>>
>>> But my ears are not good
>>> ==John ff
>>>
>>>> Hi All,
>>>>
>>>> I think I found somewhat of a showstopper bug.  I am testing using
>>>> the
>>>> latest Csound code from CVS and in rendernig this in realtime it
>>>> cuts
>>>> off the last bit of the last note with a pop.  If I render this to
>>>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts off
>>>> fairly
>>>> early in the last note; if I use PortAudio, it cuts off near the end
>>>> of the note.
>>>>
>>>> The score renders to completion with no errors reported. My guess is
>>>> that Csound isn't waiting for the audio to flush before exiting,
>>>> but I
>>>> am just guessing.  Does anyone have any ideas?  Could someone try
>>>> this
>>>> on other platforms to see if the problem occurs there too?
>>>>
>>>> Thanks,
>>>> steven
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently
>>>> attracts the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues
>>>> through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts
>>> the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts
>> the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important
>> issues through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>



-- 


Andrés

Hi Andres,

Thanks for going through the steps you went through.  I'm pretty
concerned about this.  I just started looking at Csound code but am
just getting re-familiarized with Engine code so may take me a while.
Not yet sure what is going on.

steven

On Tue, Jan 19, 2010 at 4:06 AM, Andres Cabrera  wrote:
> Hi,
>
> I can confirm this problem on Linux. I ran through QuteCsound but
> using the terminal, to make sure it was not an problem with
> QuteCsound. I used Jack to connect the output of Csound to Audacity,
> there were no xruns and I used this csd:
>
> 
> 
> 
> 
>
> sr = 44100
> ksmps = 128
> nchnls = 2
> 0dbfs = 1
>
> instr 2
> aamp linen 0.2, 0.02, p3, 0.02
> aout oscil aamp, 440, 1
> outs aout, aout
> endin
>
> 
> 
> f 1 0 1024 10 1
> i 2 0 2
> i 2 3 2
> e
> 
> 
>
> The tails of the two notes are shown in the images. It would seem that
> about 5ms were cut off. I was using -b1024 -B2048 and jack buffer size
> of 2048.
>
> Cheers,
> Andrés
>
>
> On Tue, Jan 19, 2010 at 8:58 AM, Victor Lazzarini
>  wrote:
>> Same with floats version, no audible problem, disk and dac versions
>> are the same.
>> On 19 Jan 2010, at 08:45, Victor Lazzarini wrote:
>>
>>> Same here OSX with doubles version (I have not yet built a floats
>>> one) (I hear no problems). I'll build the floats and test.
>>>
>>> Victor
>>> On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:
>>>
>>>> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using Alsa and
>>>> Portaudio
>>>>
>>>> But my ears are not good
>>>> ==John ff
>>>>
>>>>> Hi All,
>>>>>
>>>>> I think I found somewhat of a showstopper bug.  I am testing using
>>>>> the
>>>>> latest Csound code from CVS and in rendernig this in realtime it
>>>>> cuts
>>>>> off the last bit of the last note with a pop.  If I render this to
>>>>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts off
>>>>> fairly
>>>>> early in the last note; if I use PortAudio, it cuts off near the end
>>>>> of the note.
>>>>>
>>>>> The score renders to completion with no errors reported. My guess is
>>>>> that Csound isn't waiting for the audio to flush before exiting,
>>>>> but I
>>>>> am just guessing.  Does anyone have any ideas?  Could someone try
>>>>> this
>>>>> on other platforms to see if the problem occurs there too?
>>>>>
>>>>> Thanks,
>>>>> steven
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues
>>>>> through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently attracts
>>>> the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts
>>> the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts the
>> world's best and brightest in the field, creating opportunities for Conference
>> attendees to learn about information security's most important issues through
>> interactions with peers, luminaries and emerging and established companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
>
>
> --
>
>
> Andrés
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I'm holding off releasing binaries until this is sorted. As I said I  
can't see a problem on OSX, but it's better to be sure.

Victor
On 20 Jan 2010, at 04:12, Steven Yi wrote:

> Hi Andres,
>
> Thanks for going through the steps you went through.  I'm pretty
> concerned about this.  I just started looking at Csound code but am
> just getting re-familiarized with Engine code so may take me a while.
> Not yet sure what is going on.
>
> steven
>
> On Tue, Jan 19, 2010 at 4:06 AM, Andres Cabrera  
>  wrote:
>> Hi,
>>
>> I can confirm this problem on Linux. I ran through QuteCsound but
>> using the terminal, to make sure it was not an problem with
>> QuteCsound. I used Jack to connect the output of Csound to Audacity,
>> there were no xruns and I used this csd:
>>
>> 
>> 
>> 
>> 
>>
>> sr = 44100
>> ksmps = 128
>> nchnls = 2
>> 0dbfs = 1
>>
>> instr 2
>> aamp linen 0.2, 0.02, p3, 0.02
>> aout oscil aamp, 440, 1
>> outs aout, aout
>> endin
>>
>> 
>> 
>> f 1 0 1024 10 1
>> i 2 0 2
>> i 2 3 2
>> e
>> 
>> 
>>
>> The tails of the two notes are shown in the images. It would seem  
>> that
>> about 5ms were cut off. I was using -b1024 -B2048 and jack buffer  
>> size
>> of 2048.
>>
>> Cheers,
>> Andrés
>>
>>
>> On Tue, Jan 19, 2010 at 8:58 AM, Victor Lazzarini
>>  wrote:
>>> Same with floats version, no audible problem, disk and dac versions
>>> are the same.
>>> On 19 Jan 2010, at 08:45, Victor Lazzarini wrote:
>>>
>>>> Same here OSX with doubles version (I have not yet built a floats
>>>> one) (I hear no problems). I'll build the floats and test.
>>>>
>>>> Victor
>>>> On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:
>>>>
>>>>> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using  
>>>>> Alsa and
>>>>> Portaudio
>>>>>
>>>>> But my ears are not good
>>>>> ==John ff
>>>>>
>>>>>> Hi All,
>>>>>>
>>>>>> I think I found somewhat of a showstopper bug.  I am testing  
>>>>>> using
>>>>>> the
>>>>>> latest Csound code from CVS and in rendernig this in realtime it
>>>>>> cuts
>>>>>> off the last bit of the last note with a pop.  If I render this  
>>>>>> to
>>>>>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts off
>>>>>> fairly
>>>>>> early in the last note; if I use PortAudio, it cuts off near  
>>>>>> the end
>>>>>> of the note.
>>>>>>
>>>>>> The score renders to completion with no errors reported. My  
>>>>>> guess is
>>>>>> that Csound isn't waiting for the audio to flush before exiting,
>>>>>> but I
>>>>>> am just guessing.  Does anyone have any ideas?  Could someone try
>>>>>> this
>>>>>> on other platforms to see if the problem occurs there too?
>>>>>>
>>>>>> Thanks,
>>>>>> steven
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts the
>>>>>> world's best and brightest in the field, creating opportunities  
>>>>>> for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues
>>>>>> through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently  
>>>>> attracts
>>>>> the
>>>>> world's best and brightest in the field, creating opportunities  
>>>>> for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently  
>>>> attracts
>>>> the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently  
>>> attracts the
>>> world's best and brightest in the field, creating opportunities  
>>> for Conference
>>> attendees to learn about information security's most important  
>>> issues through
>>> interactions with peers, luminaries and emerging and established  
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>>
>>
>> --
>>
>>
>> Andrés
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently  
>> attracts the
>> world's best and brightest in the field, creating opportunities for  
>> Conference
>> attendees to learn about information security's most important  
>> issues through
>> interactions with peers, luminaries and emerging and established  
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Even if it quantizes, it shouldn't cut off the note but rather extends
past the end.  Event quantizing also wouldn't explain how this renders
to completion to disk but not in realtime.

I'm still having trouble finding where Csound is using the realtime
audio playback hooks. (grep shows where it gets set but not used.) I
also wonder how long this problem has been here if it is showing up
for Rene in 5.11.1.  Perhaps it has to do with the change to work in
samples for time?  If so, then that means the problem would have been
around for a long while.

On Wed, Jan 20, 2010 at 9:25 AM, Victor Lazzarini
 wrote:
> Just use lower ksmps. Your events will be quantised to control period
> boundaries, that's how Csound works.
> Victor
> On 20 Jan 2010, at 13:36, Rene Djack wrote:
>
> Hi,
>
> I can see the problem in real time and also in rendering when ksmps is high.
>
> My test csd is:
> 
> 
> -Wfd -+rtaudio=null -+rtmidi=null --expression-opt
> 
> 
> sr     = 44100
> ksmps    = 1024
> nchnls    = 1
> 0dbfs    = 1
>
> instr    2
> aout    oscil    0.5, 100, 1
>     out    aout
> endin
> 
> 
> f 1 0 1024 10 1
> i 2 0 2
> i 2 3 2
> e
> 
> 
>
> Output is:
> new alloc for instr 2:
> B  0.000 ..  3.000 T  2.995 TT  2.995 M:  0.50000
> B  3.000 ..  5.000 T  4.992 TT  4.992 M:  0.50000
> Score finished in csoundPerform().
>
>
> Note 1 start at 0.000s    end at 1.996s
> Note 2 start at 2.995s    end at 4.992s
>
> Tested on csound 5.11.1 float and double, UbuntuStudio 9.10
>
> Best,
> René
>
>
>
> 2010/1/20 Victor Lazzarini 
>>
>> I'm holding off releasing binaries until this is sorted. As I said I
>> can't see a problem on OSX, but it's better to be sure.
>>
>> Victor
>> On 20 Jan 2010, at 04:12, Steven Yi wrote:
>>
>> > Hi Andres,
>> >
>> > Thanks for going through the steps you went through.  I'm pretty
>> > concerned about this.  I just started looking at Csound code but am
>> > just getting re-familiarized with Engine code so may take me a while.
>> > Not yet sure what is going on.
>> >
>> > steven
>> >
>> > On Tue, Jan 19, 2010 at 4:06 AM, Andres Cabrera
>> >  wrote:
>> >> Hi,
>> >>
>> >> I can confirm this problem on Linux. I ran through QuteCsound but
>> >> using the terminal, to make sure it was not an problem with
>> >> QuteCsound. I used Jack to connect the output of Csound to Audacity,
>> >> there were no xruns and I used this csd:
>> >>
>> >> 
>> >> 
>> >> 
>> >> 
>> >>
>> >> sr = 44100
>> >> ksmps = 128
>> >> nchnls = 2
>> >> 0dbfs = 1
>> >>
>> >> instr 2
>> >> aamp linen 0.2, 0.02, p3, 0.02
>> >> aout oscil aamp, 440, 1
>> >> outs aout, aout
>> >> endin
>> >>
>> >> 
>> >> 
>> >> f 1 0 1024 10 1
>> >> i 2 0 2
>> >> i 2 3 2
>> >> e
>> >> 
>> >> 
>> >>
>> >> The tails of the two notes are shown in the images. It would seem
>> >> that
>> >> about 5ms were cut off. I was using -b1024 -B2048 and jack buffer
>> >> size
>> >> of 2048.
>> >>
>> >> Cheers,
>> >> Andrés
>> >>
>> >>
>> >> On Tue, Jan 19, 2010 at 8:58 AM, Victor Lazzarini
>> >>  wrote:
>> >>> Same with floats version, no audible problem, disk and dac versions
>> >>> are the same.
>> >>> On 19 Jan 2010, at 08:45, Victor Lazzarini wrote:
>> >>>
>> >>>> Same here OSX with doubles version (I have not yet built a floats
>> >>>> one) (I hear no problems). I'll build the floats and test.
>> >>>>
>> >>>> Victor
>> >>>> On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:
>> >>>>
>> >>>>> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using
>> >>>>> Alsa and
>> >>>>> Portaudio
>> >>>>>
>> >>>>> But my ears are not good
>> >>>>> ==John ff
>> >>>>>
>> >>>>>> Hi All,
>> >>>>>>
>> >>>>>> I think I found somewhat of a showstopper bug.  I am testing
>> >>>>>> using
>> >>>>>> the
>> >>>>>> latest Csound code from CVS and in rendernig this in realtime it
>> >>>>>> cuts
>> >>>>>> off the last bit of the last note with a pop.  If I render this
>> >>>>>> to
>> >>>>>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts off
>> >>>>>> fairly
>> >>>>>> early in the last note; if I use PortAudio, it cuts off near
>> >>>>>> the end
>> >>>>>> of the note.
>> >>>>>>
>> >>>>>> The score renders to completion with no errors reported. My
>> >>>>>> guess is
>> >>>>>> that Csound isn't waiting for the audio to flush before exiting,
>> >>>>>> but I
>> >>>>>> am just guessing.  Does anyone have any ideas?  Could someone try
>> >>>>>> this
>> >>>>>> on other platforms to see if the problem occurs there too?
>> >>>>>>
>> >>>>>> Thanks,
>> >>>>>> steven
>> >>>>>>
>> >>>>>> ------------------------------------------------------------------------------
>> >>>>>> Throughout its 18-year history, RSA Conference consistently
>> >>>>>> attracts the
>> >>>>>> world's best and brightest in the field, creating opportunities
>> >>>>>> for
>> >>>>>> Conference
>> >>>>>> attendees to learn about information security's most important
>> >>>>>> issues
>> >>>>>> through
>> >>>>>> interactions with peers, luminaries and emerging and established
>> >>>>>> companies.
>> >>>>>>
>> >>>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>> >>>>>> Csound-devel mailing list
>> >>>>>> Csound-devel@lists.sourceforge.net
>> >>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >>>>>>
>> >>>>>
>> >>>>>
>> >>>>>
>> >>>>>
>> >>>>> ------------------------------------------------------------------------------
>> >>>>> Throughout its 18-year history, RSA Conference consistently
>> >>>>> attracts
>> >>>>> the
>> >>>>> world's best and brightest in the field, creating opportunities
>> >>>>> for
>> >>>>> Conference
>> >>>>> attendees to learn about information security's most important
>> >>>>> issues through
>> >>>>> interactions with peers, luminaries and emerging and established
>> >>>>> companies.
>> >>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>> >>>>> _______________________________________________
>> >>>>> Csound-devel mailing list
>> >>>>> Csound-devel@lists.sourceforge.net
>> >>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >>>>
>> >>>>
>> >>>>
>> >>>> ------------------------------------------------------------------------------
>> >>>> Throughout its 18-year history, RSA Conference consistently
>> >>>> attracts
>> >>>> the
>> >>>> world's best and brightest in the field, creating opportunities for
>> >>>> Conference
>> >>>> attendees to learn about information security's most important
>> >>>> issues through
>> >>>> interactions with peers, luminaries and emerging and established
>> >>>> companies.
>> >>>> http://p.sf.net/sfu/rsaconf-dev2dev
>> >>>> _______________________________________________
>> >>>> Csound-devel mailing list
>> >>>> Csound-devel@lists.sourceforge.net
>> >>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >>>
>> >>>
>> >>>
>> >>> ------------------------------------------------------------------------------
>> >>> Throughout its 18-year history, RSA Conference consistently
>> >>> attracts the
>> >>> world's best and brightest in the field, creating opportunities
>> >>> for Conference
>> >>> attendees to learn about information security's most important
>> >>> issues through
>> >>> interactions with peers, luminaries and emerging and established
>> >>> companies.
>> >>> http://p.sf.net/sfu/rsaconf-dev2dev
>> >>> _______________________________________________
>> >>> Csound-devel mailing list
>> >>> Csound-devel@lists.sourceforge.net
>> >>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >>>
>> >>
>> >>
>> >>
>> >> --
>> >>
>> >>
>> >> Andrés
>> >>
>> >>
>> >> ------------------------------------------------------------------------------
>> >> Throughout its 18-year history, RSA Conference consistently
>> >> attracts the
>> >> world's best and brightest in the field, creating opportunities for
>> >> Conference
>> >> attendees to learn about information security's most important
>> >> issues through
>> >> interactions with peers, luminaries and emerging and established
>> >> companies.
>> >> http://p.sf.net/sfu/rsaconf-dev2dev
>> >> _______________________________________________
>> >> Csound-devel mailing list
>> >> Csound-devel@lists.sourceforge.net
>> >> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >>
>> >>
>> >
>> >
>> > ------------------------------------------------------------------------------
>> > Throughout its 18-year history, RSA Conference consistently attracts
>> > the
>> > world's best and brightest in the field, creating opportunities for
>> > Conference
>> > attendees to learn about information security's most important
>> > issues through
>> > interactions with peers, luminaries and emerging and established
>> > companies.
>> > http://p.sf.net/sfu/rsaconf-dev2dev
>> > _______________________________________________
>> > Csound-devel mailing list
>> > Csound-devel@lists.sourceforge.net
>> > https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important issues
>> through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for
> Conference
> attendees to learn about information security's most important issues
> through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for
> Conference
> attendees to learn about information security's most important issues
> through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Steven,

my comment was only to do with the post I replied to. What you report  
is a separate issue and very much a bug.
I can't reproduce it on OSX, though. But I will wait until this issue  
is fixed.
I guess if I have some time tomorrow or Friday I will build Csound on  
Linux to check.

Victor



On 20 Jan 2010, at 19:19, Steven Yi wrote:

> Even if it quantizes, it shouldn't cut off the note but rather extends
> past the end.  Event quantizing also wouldn't explain how this renders
> to completion to disk but not in realtime.
>
> I'm still having trouble finding where Csound is using the realtime
> audio playback hooks. (grep shows where it gets set but not used.) I
> also wonder how long this problem has been here if it is showing up
> for Rene in 5.11.1.  Perhaps it has to do with the change to work in
> samples for time?  If so, then that means the problem would have been
> around for a long while.
>
> On Wed, Jan 20, 2010 at 9:25 AM, Victor Lazzarini
>  wrote:
>> Just use lower ksmps. Your events will be quantised to control period
>> boundaries, that's how Csound works.
>> Victor
>> On 20 Jan 2010, at 13:36, Rene Djack wrote:
>>
>> Hi,
>>
>> I can see the problem in real time and also in rendering when ksmps  
>> is high.
>>
>> My test csd is:
>> 
>> 
>> -Wfd -+rtaudio=null -+rtmidi=null --expression-opt
>> 
>> 
>> sr     = 44100
>> ksmps    = 1024
>> nchnls    = 1
>> 0dbfs    = 1
>>
>> instr    2
>> aout    oscil    0.5, 100, 1
>>     out    aout
>> endin
>> 
>> 
>> f 1 0 1024 10 1
>> i 2 0 2
>> i 2 3 2
>> e
>> 
>> 
>>
>> Output is:
>> new alloc for instr 2:
>> B  0.000 ..  3.000 T  2.995 TT  2.995 M:  0.50000
>> B  3.000 ..  5.000 T  4.992 TT  4.992 M:  0.50000
>> Score finished in csoundPerform().
>>
>>
>> Note 1 start at 0.000s    end at 1.996s
>> Note 2 start at 2.995s    end at 4.992s
>>
>> Tested on csound 5.11.1 float and double, UbuntuStudio 9.10
>>
>> Best,
>> René
>>
>>
>>
>> 2010/1/20 Victor Lazzarini 
>>>
>>> I'm holding off releasing binaries until this is sorted. As I said I
>>> can't see a problem on OSX, but it's better to be sure.
>>>
>>> Victor
>>> On 20 Jan 2010, at 04:12, Steven Yi wrote:
>>>
>>>> Hi Andres,
>>>>
>>>> Thanks for going through the steps you went through.  I'm pretty
>>>> concerned about this.  I just started looking at Csound code but am
>>>> just getting re-familiarized with Engine code so may take me a  
>>>> while.
>>>> Not yet sure what is going on.
>>>>
>>>> steven
>>>>
>>>> On Tue, Jan 19, 2010 at 4:06 AM, Andres Cabrera
>>>>  wrote:
>>>>> Hi,
>>>>>
>>>>> I can confirm this problem on Linux. I ran through QuteCsound but
>>>>> using the terminal, to make sure it was not an problem with
>>>>> QuteCsound. I used Jack to connect the output of Csound to  
>>>>> Audacity,
>>>>> there were no xruns and I used this csd:
>>>>>
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>>>>
>>>>> sr = 44100
>>>>> ksmps = 128
>>>>> nchnls = 2
>>>>> 0dbfs = 1
>>>>>
>>>>> instr 2
>>>>> aamp linen 0.2, 0.02, p3, 0.02
>>>>> aout oscil aamp, 440, 1
>>>>> outs aout, aout
>>>>> endin
>>>>>
>>>>> 
>>>>> 
>>>>> f 1 0 1024 10 1
>>>>> i 2 0 2
>>>>> i 2 3 2
>>>>> e
>>>>> 
>>>>> 
>>>>>
>>>>> The tails of the two notes are shown in the images. It would seem
>>>>> that
>>>>> about 5ms were cut off. I was using -b1024 -B2048 and jack buffer
>>>>> size
>>>>> of 2048.
>>>>>
>>>>> Cheers,
>>>>> Andrés
>>>>>
>>>>>
>>>>> On Tue, Jan 19, 2010 at 8:58 AM, Victor Lazzarini
>>>>>  wrote:
>>>>>> Same with floats version, no audible problem, disk and dac  
>>>>>> versions
>>>>>> are the same.
>>>>>> On 19 Jan 2010, at 08:45, Victor Lazzarini wrote:
>>>>>>
>>>>>>> Same here OSX with doubles version (I have not yet built a  
>>>>>>> floats
>>>>>>> one) (I hear no problems). I'll build the floats and test.
>>>>>>>
>>>>>>> Victor
>>>>>>> On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:
>>>>>>>
>>>>>>>> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using
>>>>>>>> Alsa and
>>>>>>>> Portaudio
>>>>>>>>
>>>>>>>> But my ears are not good
>>>>>>>> ==John ff
>>>>>>>>
>>>>>>>>> Hi All,
>>>>>>>>>
>>>>>>>>> I think I found somewhat of a showstopper bug.  I am testing
>>>>>>>>> using
>>>>>>>>> the
>>>>>>>>> latest Csound code from CVS and in rendernig this in  
>>>>>>>>> realtime it
>>>>>>>>> cuts
>>>>>>>>> off the last bit of the last note with a pop.  If I render  
>>>>>>>>> this
>>>>>>>>> to
>>>>>>>>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts  
>>>>>>>>> off
>>>>>>>>> fairly
>>>>>>>>> early in the last note; if I use PortAudio, it cuts off near
>>>>>>>>> the end
>>>>>>>>> of the note.
>>>>>>>>>
>>>>>>>>> The score renders to completion with no errors reported. My
>>>>>>>>> guess is
>>>>>>>>> that Csound isn't waiting for the audio to flush before  
>>>>>>>>> exiting,
>>>>>>>>> but I
>>>>>>>>> am just guessing.  Does anyone have any ideas?  Could  
>>>>>>>>> someone try
>>>>>>>>> this
>>>>>>>>> on other platforms to see if the problem occurs there too?
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>> steven
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>> attracts the
>>>>>>>>> world's best and brightest in the field, creating  
>>>>>>>>> opportunities
>>>>>>>>> for
>>>>>>>>> Conference
>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>> issues
>>>>>>>>> through
>>>>>>>>> interactions with peers, luminaries and emerging and  
>>>>>>>>> established
>>>>>>>>> companies.
>>>>>>>>>
>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>>>>>>>> Csound-devel mailing list
>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>> attracts
>>>>>>>> the
>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>> for
>>>>>>>> Conference
>>>>>>>> attendees to learn about information security's most important
>>>>>>>> issues through
>>>>>>>> interactions with peers, luminaries and emerging and  
>>>>>>>> established
>>>>>>>> companies.
>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>> _______________________________________________
>>>>>>>> Csound-devel mailing list
>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts
>>>>>>> the
>>>>>>> world's best and brightest in the field, creating  
>>>>>>> opportunities for
>>>>>>> Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts the
>>>>>> world's best and brightest in the field, creating opportunities
>>>>>> for Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>>
>>>>>
>>>>> Andrés
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities  
>>>>> for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently  
>>>> attracts
>>>> the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently  
>>> attracts the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important  
>>> issues
>>> through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>> < 
>> ksmps 
>> .csd 
>> > 
>> < 
>> End_note1 
>> .jpg 
>> > 
>> < 
>> End_note2 
>> .jpg 
>> > 
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently  
>> attracts the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important issues
>> through
>> interactions with peers, luminaries and emerging and established  
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently  
>> attracts the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important issues
>> through
>> interactions with peers, luminaries and emerging and established  
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hi Victor,

Sounds good.  Wondering about OSX: is the driver callback-based or
blocking? I checked and using -+rtaudio=pa_cb I get the clipped sound
on Windows XP but with -+rtaudio=pa_bl I do not.  I'm hoping this will
help narrow down the problem.

What driver did you use for OSX?  Could you perhaps try with pa_bl,
pa_cb, and coreaudio if you have a chance?

Thanks!
steven

On Wed, Jan 20, 2010 at 2:29 PM, Victor Lazzarini
 wrote:
> Steven,
>
> my comment was only to do with the post I replied to. What you report
> is a separate issue and very much a bug.
> I can't reproduce it on OSX, though. But I will wait until this issue
> is fixed.
> I guess if I have some time tomorrow or Friday I will build Csound on
> Linux to check.
>
> Victor
>
>
>
> On 20 Jan 2010, at 19:19, Steven Yi wrote:
>
>> Even if it quantizes, it shouldn't cut off the note but rather extends
>> past the end.  Event quantizing also wouldn't explain how this renders
>> to completion to disk but not in realtime.
>>
>> I'm still having trouble finding where Csound is using the realtime
>> audio playback hooks. (grep shows where it gets set but not used.) I
>> also wonder how long this problem has been here if it is showing up
>> for Rene in 5.11.1.  Perhaps it has to do with the change to work in
>> samples for time?  If so, then that means the problem would have been
>> around for a long while.
>>
>> On Wed, Jan 20, 2010 at 9:25 AM, Victor Lazzarini
>>  wrote:
>>> Just use lower ksmps. Your events will be quantised to control period
>>> boundaries, that's how Csound works.
>>> Victor
>>> On 20 Jan 2010, at 13:36, Rene Djack wrote:
>>>
>>> Hi,
>>>
>>> I can see the problem in real time and also in rendering when ksmps
>>> is high.
>>>
>>> My test csd is:
>>> 
>>> 
>>> -Wfd -+rtaudio=null -+rtmidi=null --expression-opt
>>> 
>>> 
>>> sr     = 44100
>>> ksmps    = 1024
>>> nchnls    = 1
>>> 0dbfs    = 1
>>>
>>> instr    2
>>> aout    oscil    0.5, 100, 1
>>>     out    aout
>>> endin
>>> 
>>> 
>>> f 1 0 1024 10 1
>>> i 2 0 2
>>> i 2 3 2
>>> e
>>> 
>>> 
>>>
>>> Output is:
>>> new alloc for instr 2:
>>> B  0.000 ..  3.000 T  2.995 TT  2.995 M:  0.50000
>>> B  3.000 ..  5.000 T  4.992 TT  4.992 M:  0.50000
>>> Score finished in csoundPerform().
>>>
>>>
>>> Note 1 start at 0.000s    end at 1.996s
>>> Note 2 start at 2.995s    end at 4.992s
>>>
>>> Tested on csound 5.11.1 float and double, UbuntuStudio 9.10
>>>
>>> Best,
>>> René
>>>
>>>
>>>
>>> 2010/1/20 Victor Lazzarini 
>>>>
>>>> I'm holding off releasing binaries until this is sorted. As I said I
>>>> can't see a problem on OSX, but it's better to be sure.
>>>>
>>>> Victor
>>>> On 20 Jan 2010, at 04:12, Steven Yi wrote:
>>>>
>>>>> Hi Andres,
>>>>>
>>>>> Thanks for going through the steps you went through.  I'm pretty
>>>>> concerned about this.  I just started looking at Csound code but am
>>>>> just getting re-familiarized with Engine code so may take me a
>>>>> while.
>>>>> Not yet sure what is going on.
>>>>>
>>>>> steven
>>>>>
>>>>> On Tue, Jan 19, 2010 at 4:06 AM, Andres Cabrera
>>>>>  wrote:
>>>>>> Hi,
>>>>>>
>>>>>> I can confirm this problem on Linux. I ran through QuteCsound but
>>>>>> using the terminal, to make sure it was not an problem with
>>>>>> QuteCsound. I used Jack to connect the output of Csound to
>>>>>> Audacity,
>>>>>> there were no xruns and I used this csd:
>>>>>>
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>> 
>>>>>>
>>>>>> sr = 44100
>>>>>> ksmps = 128
>>>>>> nchnls = 2
>>>>>> 0dbfs = 1
>>>>>>
>>>>>> instr 2
>>>>>> aamp linen 0.2, 0.02, p3, 0.02
>>>>>> aout oscil aamp, 440, 1
>>>>>> outs aout, aout
>>>>>> endin
>>>>>>
>>>>>> 
>>>>>> 
>>>>>> f 1 0 1024 10 1
>>>>>> i 2 0 2
>>>>>> i 2 3 2
>>>>>> e
>>>>>> 
>>>>>> 
>>>>>>
>>>>>> The tails of the two notes are shown in the images. It would seem
>>>>>> that
>>>>>> about 5ms were cut off. I was using -b1024 -B2048 and jack buffer
>>>>>> size
>>>>>> of 2048.
>>>>>>
>>>>>> Cheers,
>>>>>> Andrés
>>>>>>
>>>>>>
>>>>>> On Tue, Jan 19, 2010 at 8:58 AM, Victor Lazzarini
>>>>>>  wrote:
>>>>>>> Same with floats version, no audible problem, disk and dac
>>>>>>> versions
>>>>>>> are the same.
>>>>>>> On 19 Jan 2010, at 08:45, Victor Lazzarini wrote:
>>>>>>>
>>>>>>>> Same here OSX with doubles version (I have not yet built a
>>>>>>>> floats
>>>>>>>> one) (I hear no problems). I'll build the floats and test.
>>>>>>>>
>>>>>>>> Victor
>>>>>>>> On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:
>>>>>>>>
>>>>>>>>> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using
>>>>>>>>> Alsa and
>>>>>>>>> Portaudio
>>>>>>>>>
>>>>>>>>> But my ears are not good
>>>>>>>>> ==John ff
>>>>>>>>>
>>>>>>>>>> Hi All,
>>>>>>>>>>
>>>>>>>>>> I think I found somewhat of a showstopper bug.  I am testing
>>>>>>>>>> using
>>>>>>>>>> the
>>>>>>>>>> latest Csound code from CVS and in rendernig this in
>>>>>>>>>> realtime it
>>>>>>>>>> cuts
>>>>>>>>>> off the last bit of the last note with a pop.  If I render
>>>>>>>>>> this
>>>>>>>>>> to
>>>>>>>>>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts
>>>>>>>>>> off
>>>>>>>>>> fairly
>>>>>>>>>> early in the last note; if I use PortAudio, it cuts off near
>>>>>>>>>> the end
>>>>>>>>>> of the note.
>>>>>>>>>>
>>>>>>>>>> The score renders to completion with no errors reported. My
>>>>>>>>>> guess is
>>>>>>>>>> that Csound isn't waiting for the audio to flush before
>>>>>>>>>> exiting,
>>>>>>>>>> but I
>>>>>>>>>> am just guessing.  Does anyone have any ideas?  Could
>>>>>>>>>> someone try
>>>>>>>>>> this
>>>>>>>>>> on other platforms to see if the problem occurs there too?
>>>>>>>>>>
>>>>>>>>>> Thanks,
>>>>>>>>>> steven
>>>>>>>>>>
>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>> attracts the
>>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>>> opportunities
>>>>>>>>>> for
>>>>>>>>>> Conference
>>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>>> issues
>>>>>>>>>> through
>>>>>>>>>> interactions with peers, luminaries and emerging and
>>>>>>>>>> established
>>>>>>>>>> companies.
>>>>>>>>>>
>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>> attracts
>>>>>>>>> the
>>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>>> for
>>>>>>>>> Conference
>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>> issues through
>>>>>>>>> interactions with peers, luminaries and emerging and
>>>>>>>>> established
>>>>>>>>> companies.
>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>> _______________________________________________
>>>>>>>>> Csound-devel mailing list
>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>> attracts
>>>>>>>> the
>>>>>>>> world's best and brightest in the field, creating
>>>>>>>> opportunities for
>>>>>>>> Conference
>>>>>>>> attendees to learn about information security's most important
>>>>>>>> issues through
>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>> companies.
>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>> _______________________________________________
>>>>>>>> Csound-devel mailing list
>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts the
>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>> for Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>>
>>>>>>
>>>>>> Andrés
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts the
>>>>>> world's best and brightest in the field, creating opportunities
>>>>>> for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts
>>>>> the
>>>>> world's best and brightest in the field, creating opportunities for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently
>>>> attracts the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues
>>>> through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>> <
>>> ksmps
>>> .csd
>>> >
>>> <
>>> End_note1
>>> .jpg
>>> >
>>> <
>>> End_note2
>>> .jpg
>>> >
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently
>>> attracts the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important issues
>>> through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently
>>> attracts the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important issues
>>> through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts
>> the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important
>> issues through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

it does not happen with either  pa_cb or pa_bl, but happens with  
coreaudio. It seems the device is closed before being flushed.
Looks like it's an RT module issue then.

Victor
On 20 Jan 2010, at 19:40, Steven Yi wrote:

> Hi Victor,
>
> Sounds good.  Wondering about OSX: is the driver callback-based or
> blocking? I checked and using -+rtaudio=pa_cb I get the clipped sound
> on Windows XP but with -+rtaudio=pa_bl I do not.  I'm hoping this will
> help narrow down the problem.
>
> What driver did you use for OSX?  Could you perhaps try with pa_bl,
> pa_cb, and coreaudio if you have a chance?
>
> Thanks!
> steven
>
> On Wed, Jan 20, 2010 at 2:29 PM, Victor Lazzarini
>  wrote:
>> Steven,
>>
>> my comment was only to do with the post I replied to. What you report
>> is a separate issue and very much a bug.
>> I can't reproduce it on OSX, though. But I will wait until this issue
>> is fixed.
>> I guess if I have some time tomorrow or Friday I will build Csound on
>> Linux to check.
>>
>> Victor
>>
>>
>>
>> On 20 Jan 2010, at 19:19, Steven Yi wrote:
>>
>>> Even if it quantizes, it shouldn't cut off the note but rather  
>>> extends
>>> past the end.  Event quantizing also wouldn't explain how this  
>>> renders
>>> to completion to disk but not in realtime.
>>>
>>> I'm still having trouble finding where Csound is using the realtime
>>> audio playback hooks. (grep shows where it gets set but not used.) I
>>> also wonder how long this problem has been here if it is showing up
>>> for Rene in 5.11.1.  Perhaps it has to do with the change to work in
>>> samples for time?  If so, then that means the problem would have  
>>> been
>>> around for a long while.
>>>
>>> On Wed, Jan 20, 2010 at 9:25 AM, Victor Lazzarini
>>>  wrote:
>>>> Just use lower ksmps. Your events will be quantised to control  
>>>> period
>>>> boundaries, that's how Csound works.
>>>> Victor
>>>> On 20 Jan 2010, at 13:36, Rene Djack wrote:
>>>>
>>>> Hi,
>>>>
>>>> I can see the problem in real time and also in rendering when ksmps
>>>> is high.
>>>>
>>>> My test csd is:
>>>> 
>>>> 
>>>> -Wfd -+rtaudio=null -+rtmidi=null --expression-opt
>>>> 
>>>> 
>>>> sr     = 44100
>>>> ksmps    = 1024
>>>> nchnls    = 1
>>>> 0dbfs    = 1
>>>>
>>>> instr    2
>>>> aout    oscil    0.5, 100, 1
>>>>     out    aout
>>>> endin
>>>> 
>>>> 
>>>> f 1 0 1024 10 1
>>>> i 2 0 2
>>>> i 2 3 2
>>>> e
>>>> 
>>>> 
>>>>
>>>> Output is:
>>>> new alloc for instr 2:
>>>> B  0.000 ..  3.000 T  2.995 TT  2.995 M:  0.50000
>>>> B  3.000 ..  5.000 T  4.992 TT  4.992 M:  0.50000
>>>> Score finished in csoundPerform().
>>>>
>>>>
>>>> Note 1 start at 0.000s    end at 1.996s
>>>> Note 2 start at 2.995s    end at 4.992s
>>>>
>>>> Tested on csound 5.11.1 float and double, UbuntuStudio 9.10
>>>>
>>>> Best,
>>>> René
>>>>
>>>>
>>>>
>>>> 2010/1/20 Victor Lazzarini 
>>>>>
>>>>> I'm holding off releasing binaries until this is sorted. As I  
>>>>> said I
>>>>> can't see a problem on OSX, but it's better to be sure.
>>>>>
>>>>> Victor
>>>>> On 20 Jan 2010, at 04:12, Steven Yi wrote:
>>>>>
>>>>>> Hi Andres,
>>>>>>
>>>>>> Thanks for going through the steps you went through.  I'm pretty
>>>>>> concerned about this.  I just started looking at Csound code  
>>>>>> but am
>>>>>> just getting re-familiarized with Engine code so may take me a
>>>>>> while.
>>>>>> Not yet sure what is going on.
>>>>>>
>>>>>> steven
>>>>>>
>>>>>> On Tue, Jan 19, 2010 at 4:06 AM, Andres Cabrera
>>>>>>  wrote:
>>>>>>> Hi,
>>>>>>>
>>>>>>> I can confirm this problem on Linux. I ran through QuteCsound  
>>>>>>> but
>>>>>>> using the terminal, to make sure it was not an problem with
>>>>>>> QuteCsound. I used Jack to connect the output of Csound to
>>>>>>> Audacity,
>>>>>>> there were no xruns and I used this csd:
>>>>>>>
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>>
>>>>>>> sr = 44100
>>>>>>> ksmps = 128
>>>>>>> nchnls = 2
>>>>>>> 0dbfs = 1
>>>>>>>
>>>>>>> instr 2
>>>>>>> aamp linen 0.2, 0.02, p3, 0.02
>>>>>>> aout oscil aamp, 440, 1
>>>>>>> outs aout, aout
>>>>>>> endin
>>>>>>>
>>>>>>> 
>>>>>>> 
>>>>>>> f 1 0 1024 10 1
>>>>>>> i 2 0 2
>>>>>>> i 2 3 2
>>>>>>> e
>>>>>>> 
>>>>>>> 
>>>>>>>
>>>>>>> The tails of the two notes are shown in the images. It would  
>>>>>>> seem
>>>>>>> that
>>>>>>> about 5ms were cut off. I was using -b1024 -B2048 and jack  
>>>>>>> buffer
>>>>>>> size
>>>>>>> of 2048.
>>>>>>>
>>>>>>> Cheers,
>>>>>>> Andrés
>>>>>>>
>>>>>>>
>>>>>>> On Tue, Jan 19, 2010 at 8:58 AM, Victor Lazzarini
>>>>>>>  wrote:
>>>>>>>> Same with floats version, no audible problem, disk and dac
>>>>>>>> versions
>>>>>>>> are the same.
>>>>>>>> On 19 Jan 2010, at 08:45, Victor Lazzarini wrote:
>>>>>>>>
>>>>>>>>> Same here OSX with doubles version (I have not yet built a
>>>>>>>>> floats
>>>>>>>>> one) (I hear no problems). I'll build the floats and test.
>>>>>>>>>
>>>>>>>>> Victor
>>>>>>>>> On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:
>>>>>>>>>
>>>>>>>>>> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using
>>>>>>>>>> Alsa and
>>>>>>>>>> Portaudio
>>>>>>>>>>
>>>>>>>>>> But my ears are not good
>>>>>>>>>> ==John ff
>>>>>>>>>>
>>>>>>>>>>> Hi All,
>>>>>>>>>>>
>>>>>>>>>>> I think I found somewhat of a showstopper bug.  I am testing
>>>>>>>>>>> using
>>>>>>>>>>> the
>>>>>>>>>>> latest Csound code from CVS and in rendernig this in
>>>>>>>>>>> realtime it
>>>>>>>>>>> cuts
>>>>>>>>>>> off the last bit of the last note with a pop.  If I render
>>>>>>>>>>> this
>>>>>>>>>>> to
>>>>>>>>>>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts
>>>>>>>>>>> off
>>>>>>>>>>> fairly
>>>>>>>>>>> early in the last note; if I use PortAudio, it cuts off near
>>>>>>>>>>> the end
>>>>>>>>>>> of the note.
>>>>>>>>>>>
>>>>>>>>>>> The score renders to completion with no errors reported. My
>>>>>>>>>>> guess is
>>>>>>>>>>> that Csound isn't waiting for the audio to flush before
>>>>>>>>>>> exiting,
>>>>>>>>>>> but I
>>>>>>>>>>> am just guessing.  Does anyone have any ideas?  Could
>>>>>>>>>>> someone try
>>>>>>>>>>> this
>>>>>>>>>>> on other platforms to see if the problem occurs there too?
>>>>>>>>>>>
>>>>>>>>>>> Thanks,
>>>>>>>>>>> steven
>>>>>>>>>>>
>>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>>> attracts the
>>>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>>>> opportunities
>>>>>>>>>>> for
>>>>>>>>>>> Conference
>>>>>>>>>>> attendees to learn about information security's most  
>>>>>>>>>>> important
>>>>>>>>>>> issues
>>>>>>>>>>> through
>>>>>>>>>>> interactions with peers, luminaries and emerging and
>>>>>>>>>>> established
>>>>>>>>>>> companies.
>>>>>>>>>>>
>>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>> attracts
>>>>>>>>>> the
>>>>>>>>>> world's best and brightest in the field, creating  
>>>>>>>>>> opportunities
>>>>>>>>>> for
>>>>>>>>>> Conference
>>>>>>>>>> attendees to learn about information security's most  
>>>>>>>>>> important
>>>>>>>>>> issues through
>>>>>>>>>> interactions with peers, luminaries and emerging and
>>>>>>>>>> established
>>>>>>>>>> companies.
>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>> attracts
>>>>>>>>> the
>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>> opportunities for
>>>>>>>>> Conference
>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>> issues through
>>>>>>>>> interactions with peers, luminaries and emerging and  
>>>>>>>>> established
>>>>>>>>> companies.
>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>> _______________________________________________
>>>>>>>>> Csound-devel mailing list
>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>> attracts the
>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>> for Conference
>>>>>>>> attendees to learn about information security's most important
>>>>>>>> issues through
>>>>>>>> interactions with peers, luminaries and emerging and  
>>>>>>>> established
>>>>>>>> companies.
>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>> _______________________________________________
>>>>>>>> Csound-devel mailing list
>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>>
>>>>>>>
>>>>>>> Andrés
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts the
>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>> for
>>>>>>> Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts
>>>>>> the
>>>>>> world's best and brightest in the field, creating opportunities  
>>>>>> for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities  
>>>>> for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues
>>>>> through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>> <
>>>> ksmps
>>>> .csd
>>>>>
>>>> <
>>>> End_note1
>>>> .jpg
>>>>>
>>>> <
>>>> End_note2
>>>> .jpg
>>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently
>>>> attracts the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important  
>>>> issues
>>>> through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently
>>>> attracts the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important  
>>>> issues
>>>> through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts
>>> the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently  
>> attracts the
>> world's best and brightest in the field, creating opportunities for  
>> Conference
>> attendees to learn about information security's most important  
>> issues through
>> interactions with peers, luminaries and emerging and established  
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)  
before stopping the audio callback.
In the case of pa_cb, it could be the same problem. The callback is  
being stopped before it's completely done.
Same on linux(?) I guess this is more apparent with large buffers.

Victor

On 20 Jan 2010, at 20:48, Victor Lazzarini wrote:

> it does not happen with either  pa_cb or pa_bl, but happens with
> coreaudio. It seems the device is closed before being flushed.
> Looks like it's an RT module issue then.
>
> Victor
> On 20 Jan 2010, at 19:40, Steven Yi wrote:
>
>> Hi Victor,
>>
>> Sounds good.  Wondering about OSX: is the driver callback-based or
>> blocking? I checked and using -+rtaudio=pa_cb I get the clipped sound
>> on Windows XP but with -+rtaudio=pa_bl I do not.  I'm hoping this  
>> will
>> help narrow down the problem.
>>
>> What driver did you use for OSX?  Could you perhaps try with pa_bl,
>> pa_cb, and coreaudio if you have a chance?
>>
>> Thanks!
>> steven
>>
>> On Wed, Jan 20, 2010 at 2:29 PM, Victor Lazzarini
>>  wrote:
>>> Steven,
>>>
>>> my comment was only to do with the post I replied to. What you  
>>> report
>>> is a separate issue and very much a bug.
>>> I can't reproduce it on OSX, though. But I will wait until this  
>>> issue
>>> is fixed.
>>> I guess if I have some time tomorrow or Friday I will build Csound  
>>> on
>>> Linux to check.
>>>
>>> Victor
>>>
>>>
>>>
>>> On 20 Jan 2010, at 19:19, Steven Yi wrote:
>>>
>>>> Even if it quantizes, it shouldn't cut off the note but rather
>>>> extends
>>>> past the end.  Event quantizing also wouldn't explain how this
>>>> renders
>>>> to completion to disk but not in realtime.
>>>>
>>>> I'm still having trouble finding where Csound is using the realtime
>>>> audio playback hooks. (grep shows where it gets set but not  
>>>> used.) I
>>>> also wonder how long this problem has been here if it is showing up
>>>> for Rene in 5.11.1.  Perhaps it has to do with the change to work  
>>>> in
>>>> samples for time?  If so, then that means the problem would have
>>>> been
>>>> around for a long while.
>>>>
>>>> On Wed, Jan 20, 2010 at 9:25 AM, Victor Lazzarini
>>>>  wrote:
>>>>> Just use lower ksmps. Your events will be quantised to control
>>>>> period
>>>>> boundaries, that's how Csound works.
>>>>> Victor
>>>>> On 20 Jan 2010, at 13:36, Rene Djack wrote:
>>>>>
>>>>> Hi,
>>>>>
>>>>> I can see the problem in real time and also in rendering when  
>>>>> ksmps
>>>>> is high.
>>>>>
>>>>> My test csd is:
>>>>> 
>>>>> 
>>>>> -Wfd -+rtaudio=null -+rtmidi=null --expression-opt
>>>>> 
>>>>> 
>>>>> sr     = 44100
>>>>> ksmps    = 1024
>>>>> nchnls    = 1
>>>>> 0dbfs    = 1
>>>>>
>>>>> instr    2
>>>>> aout    oscil    0.5, 100, 1
>>>>>    out    aout
>>>>> endin
>>>>> 
>>>>> 
>>>>> f 1 0 1024 10 1
>>>>> i 2 0 2
>>>>> i 2 3 2
>>>>> e
>>>>> 
>>>>> 
>>>>>
>>>>> Output is:
>>>>> new alloc for instr 2:
>>>>> B  0.000 ..  3.000 T  2.995 TT  2.995 M:  0.50000
>>>>> B  3.000 ..  5.000 T  4.992 TT  4.992 M:  0.50000
>>>>> Score finished in csoundPerform().
>>>>>
>>>>>
>>>>> Note 1 start at 0.000s    end at 1.996s
>>>>> Note 2 start at 2.995s    end at 4.992s
>>>>>
>>>>> Tested on csound 5.11.1 float and double, UbuntuStudio 9.10
>>>>>
>>>>> Best,
>>>>> René
>>>>>
>>>>>
>>>>>
>>>>> 2010/1/20 Victor Lazzarini 
>>>>>>
>>>>>> I'm holding off releasing binaries until this is sorted. As I
>>>>>> said I
>>>>>> can't see a problem on OSX, but it's better to be sure.
>>>>>>
>>>>>> Victor
>>>>>> On 20 Jan 2010, at 04:12, Steven Yi wrote:
>>>>>>
>>>>>>> Hi Andres,
>>>>>>>
>>>>>>> Thanks for going through the steps you went through.  I'm pretty
>>>>>>> concerned about this.  I just started looking at Csound code
>>>>>>> but am
>>>>>>> just getting re-familiarized with Engine code so may take me a
>>>>>>> while.
>>>>>>> Not yet sure what is going on.
>>>>>>>
>>>>>>> steven
>>>>>>>
>>>>>>> On Tue, Jan 19, 2010 at 4:06 AM, Andres Cabrera
>>>>>>>  wrote:
>>>>>>>> Hi,
>>>>>>>>
>>>>>>>> I can confirm this problem on Linux. I ran through QuteCsound
>>>>>>>> but
>>>>>>>> using the terminal, to make sure it was not an problem with
>>>>>>>> QuteCsound. I used Jack to connect the output of Csound to
>>>>>>>> Audacity,
>>>>>>>> there were no xruns and I used this csd:
>>>>>>>>
>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>>>>>>
>>>>>>>> sr = 44100
>>>>>>>> ksmps = 128
>>>>>>>> nchnls = 2
>>>>>>>> 0dbfs = 1
>>>>>>>>
>>>>>>>> instr 2
>>>>>>>> aamp linen 0.2, 0.02, p3, 0.02
>>>>>>>> aout oscil aamp, 440, 1
>>>>>>>> outs aout, aout
>>>>>>>> endin
>>>>>>>>
>>>>>>>> 
>>>>>>>> 
>>>>>>>> f 1 0 1024 10 1
>>>>>>>> i 2 0 2
>>>>>>>> i 2 3 2
>>>>>>>> e
>>>>>>>> 
>>>>>>>> 
>>>>>>>>
>>>>>>>> The tails of the two notes are shown in the images. It would
>>>>>>>> seem
>>>>>>>> that
>>>>>>>> about 5ms were cut off. I was using -b1024 -B2048 and jack
>>>>>>>> buffer
>>>>>>>> size
>>>>>>>> of 2048.
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>> Andrés
>>>>>>>>
>>>>>>>>
>>>>>>>> On Tue, Jan 19, 2010 at 8:58 AM, Victor Lazzarini
>>>>>>>>  wrote:
>>>>>>>>> Same with floats version, no audible problem, disk and dac
>>>>>>>>> versions
>>>>>>>>> are the same.
>>>>>>>>> On 19 Jan 2010, at 08:45, Victor Lazzarini wrote:
>>>>>>>>>
>>>>>>>>>> Same here OSX with doubles version (I have not yet built a
>>>>>>>>>> floats
>>>>>>>>>> one) (I hear no problems). I'll build the floats and test.
>>>>>>>>>>
>>>>>>>>>> Victor
>>>>>>>>>> On 19 Jan 2010, at 08:26, jpff@cs.bath.ac.uk wrote:
>>>>>>>>>>
>>>>>>>>>>> Did not hear anything wrong; Linux OpenSuSE11.1/x86_64 using
>>>>>>>>>>> Alsa and
>>>>>>>>>>> Portaudio
>>>>>>>>>>>
>>>>>>>>>>> But my ears are not good
>>>>>>>>>>> ==John ff
>>>>>>>>>>>
>>>>>>>>>>>> Hi All,
>>>>>>>>>>>>
>>>>>>>>>>>> I think I found somewhat of a showstopper bug.  I am  
>>>>>>>>>>>> testing
>>>>>>>>>>>> using
>>>>>>>>>>>> the
>>>>>>>>>>>> latest Csound code from CVS and in rendernig this in
>>>>>>>>>>>> realtime it
>>>>>>>>>>>> cuts
>>>>>>>>>>>> off the last bit of the last note with a pop.  If I render
>>>>>>>>>>>> this
>>>>>>>>>>>> to
>>>>>>>>>>>> disk it is fine.  I'm on Windows XP.  If I use MME, it cuts
>>>>>>>>>>>> off
>>>>>>>>>>>> fairly
>>>>>>>>>>>> early in the last note; if I use PortAudio, it cuts off  
>>>>>>>>>>>> near
>>>>>>>>>>>> the end
>>>>>>>>>>>> of the note.
>>>>>>>>>>>>
>>>>>>>>>>>> The score renders to completion with no errors reported. My
>>>>>>>>>>>> guess is
>>>>>>>>>>>> that Csound isn't waiting for the audio to flush before
>>>>>>>>>>>> exiting,
>>>>>>>>>>>> but I
>>>>>>>>>>>> am just guessing.  Does anyone have any ideas?  Could
>>>>>>>>>>>> someone try
>>>>>>>>>>>> this
>>>>>>>>>>>> on other platforms to see if the problem occurs there too?
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks,
>>>>>>>>>>>> steven
>>>>>>>>>>>>
>>>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>>>> attracts the
>>>>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>>>>> opportunities
>>>>>>>>>>>> for
>>>>>>>>>>>> Conference
>>>>>>>>>>>> attendees to learn about information security's most
>>>>>>>>>>>> important
>>>>>>>>>>>> issues
>>>>>>>>>>>> through
>>>>>>>>>>>> interactions with peers, luminaries and emerging and
>>>>>>>>>>>> established
>>>>>>>>>>>> companies.
>>>>>>>>>>>>
>>>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>>> attracts
>>>>>>>>>>> the
>>>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>>>> opportunities
>>>>>>>>>>> for
>>>>>>>>>>> Conference
>>>>>>>>>>> attendees to learn about information security's most
>>>>>>>>>>> important
>>>>>>>>>>> issues through
>>>>>>>>>>> interactions with peers, luminaries and emerging and
>>>>>>>>>>> established
>>>>>>>>>>> companies.
>>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>> attracts
>>>>>>>>>> the
>>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>>> opportunities for
>>>>>>>>>> Conference
>>>>>>>>>> attendees to learn about information security's most  
>>>>>>>>>> important
>>>>>>>>>> issues through
>>>>>>>>>> interactions with peers, luminaries and emerging and
>>>>>>>>>> established
>>>>>>>>>> companies.
>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>> attracts the
>>>>>>>>> world's best and brightest in the field, creating  
>>>>>>>>> opportunities
>>>>>>>>> for Conference
>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>> issues through
>>>>>>>>> interactions with peers, luminaries and emerging and
>>>>>>>>> established
>>>>>>>>> companies.
>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>> _______________________________________________
>>>>>>>>> Csound-devel mailing list
>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> --
>>>>>>>>
>>>>>>>>
>>>>>>>> Andrés
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>> attracts the
>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>> for
>>>>>>>> Conference
>>>>>>>> attendees to learn about information security's most important
>>>>>>>> issues through
>>>>>>>> interactions with peers, luminaries and emerging and  
>>>>>>>> established
>>>>>>>> companies.
>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>> _______________________________________________
>>>>>>>> Csound-devel mailing list
>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts
>>>>>>> the
>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>> for
>>>>>>> Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts the
>>>>>> world's best and brightest in the field, creating opportunities
>>>>>> for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues
>>>>>> through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>> <
>>>>> ksmps
>>>>> .csd
>>>>>>
>>>>> <
>>>>> End_note1
>>>>> .jpg
>>>>>>
>>>>> <
>>>>> End_note2
>>>>> .jpg
>>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities  
>>>>> for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues
>>>>> through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev_______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities  
>>>>> for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues
>>>>> through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently  
>>>> attracts
>>>> the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently
>>> attracts the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts
>> the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important
>> issues through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Was this in 5/.11 or has something shaken?

> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
> before stopping the audio callback.
> In the case of pa_cb, it could be the same problem. The callback is
> being stopped before it's completely done.
> Same on linux(?) I guess this is more apparent with large buffers.
>
> Victor
>




------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hi John,

Rene reported the issue in 5.11.1 for both float and double on Ubuntu
Studio 9.10.

steven

On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
> Was this in 5/.11 or has something shaken?
>
>> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
>> before stopping the audio callback.
>> In the case of pa_cb, it could be the same problem. The callback is
>> being stopped before it's completely done.
>> Same on linux(?) I guess this is more apparent with large buffers.
>>
>> Victor
>>
>
>
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I never noticed it with 5.11, but then all's well with the default IO  
module on OSX, and I had not been testing rtcoreaudio.c at all before.
Victor
On 20 Jan 2010, at 21:49, jpff@cs.bath.ac.uk wrote:

> Was this in 5/.11 or has something shaken?
>
>> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
>> before stopping the audio callback.
>> In the case of pa_cb, it could be the same problem. The callback is
>> being stopped before it's completely done.
>> Same on linux(?) I guess this is more apparent with large buffers.
>>
>> Victor
>>
>
>
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

No. I think what he reported has nothing to do with this. His test was  
with high ksmps and the problem was rendering to disk. I think the  
behaviour he reports is the expected. What you have reported is with  
normal conditions of ksmps and shows a difference between rendering to  
disk and RT.

There are no reports this was happening on 5.11. That does not mean it  
was not.

Victor

On 20 Jan 2010, at 21:51, Steven Yi wrote:

> Hi John,
>
> Rene reported the issue in 5.11.1 for both float and double on Ubuntu
> Studio 9.10.
>
> steven
>
> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>> Was this in 5/.11 or has something shaken?
>>
>>> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
>>> before stopping the audio callback.
>>> In the case of pa_cb, it could be the same problem. The callback is
>>> being stopped before it's completely done.
>>> Same on linux(?) I guess this is more apparent with large buffers.
>>>
>>> Victor
>>>
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently  
>> attracts the
>> world's best and brightest in the field, creating opportunities for  
>> Conference
>> attendees to learn about information security's most important  
>> issues through
>> interactions with peers, luminaries and emerging and established  
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I just reread Rene's email and he mentioned:

"I can see the problem in real time and also in rendering when ksmps is high."

I did not see any mention of rendering to disk anywhere.

On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
 wrote:
> No. I think what he reported has nothing to do with this. His test was
> with high ksmps and the problem was rendering to disk. I think the
> behaviour he reports is the expected. What you have reported is with
> normal conditions of ksmps and shows a difference between rendering to
> disk and RT.
>
> There are no reports this was happening on 5.11. That does not mean it
> was not.
>
> Victor
>
> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>
>> Hi John,
>>
>> Rene reported the issue in 5.11.1 for both float and double on Ubuntu
>> Studio 9.10.
>>
>> steven
>>
>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>> Was this in 5/.11 or has something shaken?
>>>
>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
>>>> before stopping the audio callback.
>>>> In the case of pa_cb, it could be the same problem. The callback is
>>>> being stopped before it's completely done.
>>>> Same on linux(?) I guess this is more apparent with large buffers.
>>>>
>>>> Victor
>>>>
>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently
>>> attracts the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts
>> the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important
>> issues through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I think rendering means 'to disk'. His CSD had a very large ksmps.  
There is no other CSD example with normal ksmps.
I understood 'the problem when ksmps is high' in 'realtime and  
rendering [to disk]' is what he meant.

Victor

On 20 Jan 2010, at 22:10, Steven Yi wrote:

> I just reread Rene's email and he mentioned:
>
> "I can see the problem in real time and also in rendering when ksmps  
> is high."
>
> I did not see any mention of rendering to disk anywhere.
>
> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>  wrote:
>> No. I think what he reported has nothing to do with this. His test  
>> was
>> with high ksmps and the problem was rendering to disk. I think the
>> behaviour he reports is the expected. What you have reported is with
>> normal conditions of ksmps and shows a difference between rendering  
>> to
>> disk and RT.
>>
>> There are no reports this was happening on 5.11. That does not mean  
>> it
>> was not.
>>
>> Victor
>>
>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>
>>> Hi John,
>>>
>>> Rene reported the issue in 5.11.1 for both float and double on  
>>> Ubuntu
>>> Studio 9.10.
>>>
>>> steven
>>>
>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>> Was this in 5/.11 or has something shaken?
>>>>
>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
>>>>> before stopping the audio callback.
>>>>> In the case of pa_cb, it could be the same problem. The callback  
>>>>> is
>>>>> being stopped before it's completely done.
>>>>> Same on linux(?) I guess this is more apparent with large buffers.
>>>>>
>>>>> Victor
>>>>>
>>>>
>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently
>>>> attracts the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts
>>> the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently  
>> attracts the
>> world's best and brightest in the field, creating opportunities for  
>> Conference
>> attendees to learn about information security's most important  
>> issues through
>> interactions with peers, luminaries and emerging and established  
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Well, even if ksmps is high, it should still render in realtime
completely through the event, and may be a clue.  I'm going to test on
fedora and ubuntu virtual machines.  I think Fedora has csound 5.10
and will use whatever is in Ubuntu (i assume 5.11.1 from Rene's
setup). I'll use the same test.csd I originally posted. Will report
back when I have results.

On Wed, Jan 20, 2010 at 5:16 PM, Victor Lazzarini
 wrote:
> I think rendering means 'to disk'. His CSD had a very large ksmps.
> There is no other CSD example with normal ksmps.
> I understood 'the problem when ksmps is high' in 'realtime and
> rendering [to disk]' is what he meant.
>
> Victor
>
> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>
>> I just reread Rene's email and he mentioned:
>>
>> "I can see the problem in real time and also in rendering when ksmps
>> is high."
>>
>> I did not see any mention of rendering to disk anywhere.
>>
>> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>>  wrote:
>>> No. I think what he reported has nothing to do with this. His test
>>> was
>>> with high ksmps and the problem was rendering to disk. I think the
>>> behaviour he reports is the expected. What you have reported is with
>>> normal conditions of ksmps and shows a difference between rendering
>>> to
>>> disk and RT.
>>>
>>> There are no reports this was happening on 5.11. That does not mean
>>> it
>>> was not.
>>>
>>> Victor
>>>
>>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>>
>>>> Hi John,
>>>>
>>>> Rene reported the issue in 5.11.1 for both float and double on
>>>> Ubuntu
>>>> Studio 9.10.
>>>>
>>>> steven
>>>>
>>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>>> Was this in 5/.11 or has something shaken?
>>>>>
>>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
>>>>>> before stopping the audio callback.
>>>>>> In the case of pa_cb, it could be the same problem. The callback
>>>>>> is
>>>>>> being stopped before it's completely done.
>>>>>> Same on linux(?) I guess this is more apparent with large buffers.
>>>>>>
>>>>>> Victor
>>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently attracts
>>>> the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently
>>> attracts the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts
>> the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important
>> issues through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

No worries, I think I misunderstood what Victor was getting at as
well, as I didn't realize the test csd didn't have an envelope on the
oscillator. I appreciate your contribution to this thread, for what
it's worth.

On Wed, Jan 20, 2010 at 5:46 PM, Rene Djack  wrote:
> YES
>
> I have included my test csd in my previous message
>
>  -Wfd -+rtaudio=null -+rtmidi=null --expression-opt 
>
> is rendering to disk.
>
> Of course my test ksmps is very high, it  was just to show better the
> "problem".
> In QuteCsound i use high ksmps for instrument reading the gui
> (ksmps=256 for SteamPipe.csd i recently post in user list)
>
> Anyway you say it is normal, that's all.
>
> Sorry for the the noise, it's my last message on this dev-list.
>
> René
>
>
>
> 2010/1/20 Victor Lazzarini 
>>
>> I think rendering means 'to disk'. His CSD had a very large ksmps.
>> There is no other CSD example with normal ksmps.
>> I understood 'the problem when ksmps is high' in 'realtime and
>> rendering [to disk]' is what he meant.
>>
>> Victor
>>
>> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>>
>> > I just reread Rene's email and he mentioned:
>> >
>> > "I can see the problem in real time and also in rendering when ksmps
>> > is high."
>> >
>> > I did not see any mention of rendering to disk anywhere.
>> >
>> > On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>> >  wrote:
>> >> No. I think what he reported has nothing to do with this. His test
>> >> was
>> >> with high ksmps and the problem was rendering to disk. I think the
>> >> behaviour he reports is the expected. What you have reported is with
>> >> normal conditions of ksmps and shows a difference between rendering
>> >> to
>> >> disk and RT.
>> >>
>> >> There are no reports this was happening on 5.11. That does not mean
>> >> it
>> >> was not.
>> >>
>> >> Victor
>> >>
>> >> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>> >>
>> >>> Hi John,
>> >>>
>> >>> Rene reported the issue in 5.11.1 for both float and double on
>> >>> Ubuntu
>> >>> Studio 9.10.
>> >>>
>> >>> steven
>> >>>
>> >>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>> >>>> Was this in 5/.11 or has something shaken?
>> >>>>
>> >>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
>> >>>>> before stopping the audio callback.
>> >>>>> In the case of pa_cb, it could be the same problem. The callback
>> >>>>> is
>> >>>>> being stopped before it's completely done.
>> >>>>> Same on linux(?) I guess this is more apparent with large buffers.
>> >>>>>
>> >>>>> Victor
>> >>>>>
>> >>>>
>> >>>>
>> >>>>
>> >>>>
>> >>>>
>> >>>> ------------------------------------------------------------------------------
>> >>>> Throughout its 18-year history, RSA Conference consistently
>> >>>> attracts the
>> >>>> world's best and brightest in the field, creating opportunities for
>> >>>> Conference
>> >>>> attendees to learn about information security's most important
>> >>>> issues through
>> >>>> interactions with peers, luminaries and emerging and established
>> >>>> companies.
>> >>>> http://p.sf.net/sfu/rsaconf-dev2dev
>> >>>> _______________________________________________
>> >>>> Csound-devel mailing list
>> >>>> Csound-devel@lists.sourceforge.net
>> >>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >>>>
>> >>>
>> >>>
>> >>> ------------------------------------------------------------------------------
>> >>> Throughout its 18-year history, RSA Conference consistently attracts
>> >>> the
>> >>> world's best and brightest in the field, creating opportunities for
>> >>> Conference
>> >>> attendees to learn about information security's most important
>> >>> issues through
>> >>> interactions with peers, luminaries and emerging and established
>> >>> companies.
>> >>> http://p.sf.net/sfu/rsaconf-dev2dev
>> >>> _______________________________________________
>> >>> Csound-devel mailing list
>> >>> Csound-devel@lists.sourceforge.net
>> >>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >>
>> >>
>> >>
>> >> ------------------------------------------------------------------------------
>> >> Throughout its 18-year history, RSA Conference consistently
>> >> attracts the
>> >> world's best and brightest in the field, creating opportunities for
>> >> Conference
>> >> attendees to learn about information security's most important
>> >> issues through
>> >> interactions with peers, luminaries and emerging and established
>> >> companies.
>> >> http://p.sf.net/sfu/rsaconf-dev2dev
>> >> _______________________________________________
>> >> Csound-devel mailing list
>> >> Csound-devel@lists.sourceforge.net
>> >> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >>
>> >
>> >
>> > ------------------------------------------------------------------------------
>> > Throughout its 18-year history, RSA Conference consistently attracts
>> > the
>> > world's best and brightest in the field, creating opportunities for
>> > Conference
>> > attendees to learn about information security's most important
>> > issues through
>> > interactions with peers, luminaries and emerging and established
>> > companies.
>> > http://p.sf.net/sfu/rsaconf-dev2dev
>> > _______________________________________________
>> > Csound-devel mailing list
>> > Csound-devel@lists.sourceforge.net
>> > https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important issues
>> through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for
> Conference
> attendees to learn about information security's most important issues
> through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Okay I just tested using:

Csound version 5.10 beta (float samples) Dec  3 2009

This is on Fedora 12, using what is in the RPM repository which is:

csound-5.10.1-14.fc12.i686

I tested with alsa using -b1024 -B4096 and heard the premature click.
I ran it using pulseaudio and heard the click.  It seems to be not
consistent to my ear how much of a click, which makes me think there's
thread timing involved where more of the audio may or may not get
thorough before closing.

In addition, I rendered it to disk and played using the gnome media
player and it did not have a click.

Note, the build is a float build. (My tests on WinXP are with a doubles build).

Going to test on Ubuntu now.

steven

On Wed, Jan 20, 2010 at 5:33 PM, Steven Yi  wrote:
> Well, even if ksmps is high, it should still render in realtime
> completely through the event, and may be a clue.  I'm going to test on
> fedora and ubuntu virtual machines.  I think Fedora has csound 5.10
> and will use whatever is in Ubuntu (i assume 5.11.1 from Rene's
> setup). I'll use the same test.csd I originally posted. Will report
> back when I have results.
>
> On Wed, Jan 20, 2010 at 5:16 PM, Victor Lazzarini
>  wrote:
>> I think rendering means 'to disk'. His CSD had a very large ksmps.
>> There is no other CSD example with normal ksmps.
>> I understood 'the problem when ksmps is high' in 'realtime and
>> rendering [to disk]' is what he meant.
>>
>> Victor
>>
>> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>>
>>> I just reread Rene's email and he mentioned:
>>>
>>> "I can see the problem in real time and also in rendering when ksmps
>>> is high."
>>>
>>> I did not see any mention of rendering to disk anywhere.
>>>
>>> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>>>  wrote:
>>>> No. I think what he reported has nothing to do with this. His test
>>>> was
>>>> with high ksmps and the problem was rendering to disk. I think the
>>>> behaviour he reports is the expected. What you have reported is with
>>>> normal conditions of ksmps and shows a difference between rendering
>>>> to
>>>> disk and RT.
>>>>
>>>> There are no reports this was happening on 5.11. That does not mean
>>>> it
>>>> was not.
>>>>
>>>> Victor
>>>>
>>>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>>>
>>>>> Hi John,
>>>>>
>>>>> Rene reported the issue in 5.11.1 for both float and double on
>>>>> Ubuntu
>>>>> Studio 9.10.
>>>>>
>>>>> steven
>>>>>
>>>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>>>> Was this in 5/.11 or has something shaken?
>>>>>>
>>>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
>>>>>>> before stopping the audio callback.
>>>>>>> In the case of pa_cb, it could be the same problem. The callback
>>>>>>> is
>>>>>>> being stopped before it's completely done.
>>>>>>> Same on linux(?) I guess this is more apparent with large buffers.
>>>>>>>
>>>>>>> Victor
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts the
>>>>>> world's best and brightest in the field, creating opportunities for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently attracts
>>>>> the
>>>>> world's best and brightest in the field, creating opportunities for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently
>>>> attracts the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts
>>> the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts the
>> world's best and brightest in the field, creating opportunities for Conference
>> attendees to learn about information security's most important issues through
>> interactions with peers, luminaries and emerging and established companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Why don't  you try to stick  a sleep() or similar  call before any  
call to
Pa_StopStream(dev->handle) just to check if that changes things (as it  
does in rtcoreaudio())?

Victor
On 20 Jan 2010, at 23:17, Steven Yi wrote:

> Okay I just tested using:
>
> Csound version 5.10 beta (float samples) Dec  3 2009
>
> This is on Fedora 12, using what is in the RPM repository which is:
>
> csound-5.10.1-14.fc12.i686
>
> I tested with alsa using -b1024 -B4096 and heard the premature click.
> I ran it using pulseaudio and heard the click.  It seems to be not
> consistent to my ear how much of a click, which makes me think there's
> thread timing involved where more of the audio may or may not get
> thorough before closing.
>
> In addition, I rendered it to disk and played using the gnome media
> player and it did not have a click.
>
> Note, the build is a float build. (My tests on WinXP are with a  
> doubles build).
>
> Going to test on Ubuntu now.
>
> steven
>
> On Wed, Jan 20, 2010 at 5:33 PM, Steven Yi  wrote:
>> Well, even if ksmps is high, it should still render in realtime
>> completely through the event, and may be a clue.  I'm going to test  
>> on
>> fedora and ubuntu virtual machines.  I think Fedora has csound 5.10
>> and will use whatever is in Ubuntu (i assume 5.11.1 from Rene's
>> setup). I'll use the same test.csd I originally posted. Will report
>> back when I have results.
>>
>> On Wed, Jan 20, 2010 at 5:16 PM, Victor Lazzarini
>>  wrote:
>>> I think rendering means 'to disk'. His CSD had a very large ksmps.
>>> There is no other CSD example with normal ksmps.
>>> I understood 'the problem when ksmps is high' in 'realtime and
>>> rendering [to disk]' is what he meant.
>>>
>>> Victor
>>>
>>> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>>>
>>>> I just reread Rene's email and he mentioned:
>>>>
>>>> "I can see the problem in real time and also in rendering when  
>>>> ksmps
>>>> is high."
>>>>
>>>> I did not see any mention of rendering to disk anywhere.
>>>>
>>>> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>>>>  wrote:
>>>>> No. I think what he reported has nothing to do with this. His test
>>>>> was
>>>>> with high ksmps and the problem was rendering to disk. I think the
>>>>> behaviour he reports is the expected. What you have reported is  
>>>>> with
>>>>> normal conditions of ksmps and shows a difference between  
>>>>> rendering
>>>>> to
>>>>> disk and RT.
>>>>>
>>>>> There are no reports this was happening on 5.11. That does not  
>>>>> mean
>>>>> it
>>>>> was not.
>>>>>
>>>>> Victor
>>>>>
>>>>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>>>>
>>>>>> Hi John,
>>>>>>
>>>>>> Rene reported the issue in 5.11.1 for both float and double on
>>>>>> Ubuntu
>>>>>> Studio 9.10.
>>>>>>
>>>>>> steven
>>>>>>
>>>>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>>>>> Was this in 5/.11 or has something shaken?
>>>>>>>
>>>>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a  
>>>>>>>> sleep(1)
>>>>>>>> before stopping the audio callback.
>>>>>>>> In the case of pa_cb, it could be the same problem. The  
>>>>>>>> callback
>>>>>>>> is
>>>>>>>> being stopped before it's completely done.
>>>>>>>> Same on linux(?) I guess this is more apparent with large  
>>>>>>>> buffers.
>>>>>>>>
>>>>>>>> Victor
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts the
>>>>>>> world's best and brightest in the field, creating  
>>>>>>> opportunities for
>>>>>>> Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently  
>>>>>> attracts
>>>>>> the
>>>>>> world's best and brightest in the field, creating opportunities  
>>>>>> for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities  
>>>>> for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently  
>>>> attracts
>>>> the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently  
>>> attracts the
>>> world's best and brightest in the field, creating opportunities  
>>> for Conference
>>> attendees to learn about information security's most important  
>>> issues through
>>> interactions with peers, luminaries and emerging and established  
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

On Ubuntu, the version of Csound I have installed must not be the
latest from repos but perhaps that's better.  The reported version is:

Csound version 5.10 (double samples) Jul 29 2009

It gives issues again with pa_bl, alsa, and pulse audio.  No problems
with pa_cb (opposite of what I found on Windows I think).


On Wed, Jan 20, 2010 at 6:17 PM, Steven Yi  wrote:
> Okay I just tested using:
>
> Csound version 5.10 beta (float samples) Dec  3 2009
>
> This is on Fedora 12, using what is in the RPM repository which is:
>
> csound-5.10.1-14.fc12.i686
>
> I tested with alsa using -b1024 -B4096 and heard the premature click.
> I ran it using pulseaudio and heard the click.  It seems to be not
> consistent to my ear how much of a click, which makes me think there's
> thread timing involved where more of the audio may or may not get
> thorough before closing.
>
> In addition, I rendered it to disk and played using the gnome media
> player and it did not have a click.
>
> Note, the build is a float build. (My tests on WinXP are with a doubles build).
>
> Going to test on Ubuntu now.
>
> steven
>
> On Wed, Jan 20, 2010 at 5:33 PM, Steven Yi  wrote:
>> Well, even if ksmps is high, it should still render in realtime
>> completely through the event, and may be a clue.  I'm going to test on
>> fedora and ubuntu virtual machines.  I think Fedora has csound 5.10
>> and will use whatever is in Ubuntu (i assume 5.11.1 from Rene's
>> setup). I'll use the same test.csd I originally posted. Will report
>> back when I have results.
>>
>> On Wed, Jan 20, 2010 at 5:16 PM, Victor Lazzarini
>>  wrote:
>>> I think rendering means 'to disk'. His CSD had a very large ksmps.
>>> There is no other CSD example with normal ksmps.
>>> I understood 'the problem when ksmps is high' in 'realtime and
>>> rendering [to disk]' is what he meant.
>>>
>>> Victor
>>>
>>> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>>>
>>>> I just reread Rene's email and he mentioned:
>>>>
>>>> "I can see the problem in real time and also in rendering when ksmps
>>>> is high."
>>>>
>>>> I did not see any mention of rendering to disk anywhere.
>>>>
>>>> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>>>>  wrote:
>>>>> No. I think what he reported has nothing to do with this. His test
>>>>> was
>>>>> with high ksmps and the problem was rendering to disk. I think the
>>>>> behaviour he reports is the expected. What you have reported is with
>>>>> normal conditions of ksmps and shows a difference between rendering
>>>>> to
>>>>> disk and RT.
>>>>>
>>>>> There are no reports this was happening on 5.11. That does not mean
>>>>> it
>>>>> was not.
>>>>>
>>>>> Victor
>>>>>
>>>>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>>>>
>>>>>> Hi John,
>>>>>>
>>>>>> Rene reported the issue in 5.11.1 for both float and double on
>>>>>> Ubuntu
>>>>>> Studio 9.10.
>>>>>>
>>>>>> steven
>>>>>>
>>>>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>>>>> Was this in 5/.11 or has something shaken?
>>>>>>>
>>>>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a sleep(1)
>>>>>>>> before stopping the audio callback.
>>>>>>>> In the case of pa_cb, it could be the same problem. The callback
>>>>>>>> is
>>>>>>>> being stopped before it's completely done.
>>>>>>>> Same on linux(?) I guess this is more apparent with large buffers.
>>>>>>>>
>>>>>>>> Victor
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts the
>>>>>>> world's best and brightest in the field, creating opportunities for
>>>>>>> Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently attracts
>>>>>> the
>>>>>> world's best and brightest in the field, creating opportunities for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently attracts
>>>> the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts the
>>> world's best and brightest in the field, creating opportunities for Conference
>>> attendees to learn about information security's most important issues through
>>> interactions with peers, luminaries and emerging and established companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Well, from the looks of it, it seems a problem that is existing in
multiple drivers.  I'm wondering if the problem lies elsewhere and is
a flaw in Csound that if fixed would fix all the issue at once.  It
seems to go back to at least 5.10 from what I found on Ubuntu.  I want
to understand the interaction between Csound and the audio output a
bit more as I'd like to know exactly why this is happening. I'm
worried if we just put in a sleep that it may mask the problem and
that it may come up again elsewhere if we don't address the core
issue.  (Not to say that we may just need to do a sleep in the end,
but at the moment I just don't know what the root cause is.)

On Wed, Jan 20, 2010 at 6:44 PM, Victor Lazzarini
 wrote:
> Why don't  you try to stick  a sleep() or similar  call before any
> call to
> Pa_StopStream(dev->handle) just to check if that changes things (as it
> does in rtcoreaudio())?
>
> Victor
> On 20 Jan 2010, at 23:17, Steven Yi wrote:
>
>> Okay I just tested using:
>>
>> Csound version 5.10 beta (float samples) Dec  3 2009
>>
>> This is on Fedora 12, using what is in the RPM repository which is:
>>
>> csound-5.10.1-14.fc12.i686
>>
>> I tested with alsa using -b1024 -B4096 and heard the premature click.
>> I ran it using pulseaudio and heard the click.  It seems to be not
>> consistent to my ear how much of a click, which makes me think there's
>> thread timing involved where more of the audio may or may not get
>> thorough before closing.
>>
>> In addition, I rendered it to disk and played using the gnome media
>> player and it did not have a click.
>>
>> Note, the build is a float build. (My tests on WinXP are with a
>> doubles build).
>>
>> Going to test on Ubuntu now.
>>
>> steven
>>
>> On Wed, Jan 20, 2010 at 5:33 PM, Steven Yi  wrote:
>>> Well, even if ksmps is high, it should still render in realtime
>>> completely through the event, and may be a clue.  I'm going to test
>>> on
>>> fedora and ubuntu virtual machines.  I think Fedora has csound 5.10
>>> and will use whatever is in Ubuntu (i assume 5.11.1 from Rene's
>>> setup). I'll use the same test.csd I originally posted. Will report
>>> back when I have results.
>>>
>>> On Wed, Jan 20, 2010 at 5:16 PM, Victor Lazzarini
>>>  wrote:
>>>> I think rendering means 'to disk'. His CSD had a very large ksmps.
>>>> There is no other CSD example with normal ksmps.
>>>> I understood 'the problem when ksmps is high' in 'realtime and
>>>> rendering [to disk]' is what he meant.
>>>>
>>>> Victor
>>>>
>>>> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>>>>
>>>>> I just reread Rene's email and he mentioned:
>>>>>
>>>>> "I can see the problem in real time and also in rendering when
>>>>> ksmps
>>>>> is high."
>>>>>
>>>>> I did not see any mention of rendering to disk anywhere.
>>>>>
>>>>> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>>>>>  wrote:
>>>>>> No. I think what he reported has nothing to do with this. His test
>>>>>> was
>>>>>> with high ksmps and the problem was rendering to disk. I think the
>>>>>> behaviour he reports is the expected. What you have reported is
>>>>>> with
>>>>>> normal conditions of ksmps and shows a difference between
>>>>>> rendering
>>>>>> to
>>>>>> disk and RT.
>>>>>>
>>>>>> There are no reports this was happening on 5.11. That does not
>>>>>> mean
>>>>>> it
>>>>>> was not.
>>>>>>
>>>>>> Victor
>>>>>>
>>>>>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>>>>>
>>>>>>> Hi John,
>>>>>>>
>>>>>>> Rene reported the issue in 5.11.1 for both float and double on
>>>>>>> Ubuntu
>>>>>>> Studio 9.10.
>>>>>>>
>>>>>>> steven
>>>>>>>
>>>>>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>>>>>> Was this in 5/.11 or has something shaken?
>>>>>>>>
>>>>>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a
>>>>>>>>> sleep(1)
>>>>>>>>> before stopping the audio callback.
>>>>>>>>> In the case of pa_cb, it could be the same problem. The
>>>>>>>>> callback
>>>>>>>>> is
>>>>>>>>> being stopped before it's completely done.
>>>>>>>>> Same on linux(?) I guess this is more apparent with large
>>>>>>>>> buffers.
>>>>>>>>>
>>>>>>>>> Victor
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>> attracts the
>>>>>>>> world's best and brightest in the field, creating
>>>>>>>> opportunities for
>>>>>>>> Conference
>>>>>>>> attendees to learn about information security's most important
>>>>>>>> issues through
>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>> companies.
>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>> _______________________________________________
>>>>>>>> Csound-devel mailing list
>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts
>>>>>>> the
>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>> for
>>>>>>> Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts the
>>>>>> world's best and brightest in the field, creating opportunities
>>>>>> for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts
>>>>> the
>>>>> world's best and brightest in the field, creating opportunities for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently
>>>> attracts the
>>>> world's best and brightest in the field, creating opportunities
>>>> for Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts
>> the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important
>> issues through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Okay, I'm seeing that csound_main.c calls csoundDestroy but does not
call csoundCleanup. I'm pretty sure it is supposed to: can anyone
confirm?  Going go test this now.

On Wed, Jan 20, 2010 at 7:29 PM, Steven Yi  wrote:
> Well, from the looks of it, it seems a problem that is existing in
> multiple drivers.  I'm wondering if the problem lies elsewhere and is
> a flaw in Csound that if fixed would fix all the issue at once.  It
> seems to go back to at least 5.10 from what I found on Ubuntu.  I want
> to understand the interaction between Csound and the audio output a
> bit more as I'd like to know exactly why this is happening. I'm
> worried if we just put in a sleep that it may mask the problem and
> that it may come up again elsewhere if we don't address the core
> issue.  (Not to say that we may just need to do a sleep in the end,
> but at the moment I just don't know what the root cause is.)
>
> On Wed, Jan 20, 2010 at 6:44 PM, Victor Lazzarini
>  wrote:
>> Why don't  you try to stick  a sleep() or similar  call before any
>> call to
>> Pa_StopStream(dev->handle) just to check if that changes things (as it
>> does in rtcoreaudio())?
>>
>> Victor
>> On 20 Jan 2010, at 23:17, Steven Yi wrote:
>>
>>> Okay I just tested using:
>>>
>>> Csound version 5.10 beta (float samples) Dec  3 2009
>>>
>>> This is on Fedora 12, using what is in the RPM repository which is:
>>>
>>> csound-5.10.1-14.fc12.i686
>>>
>>> I tested with alsa using -b1024 -B4096 and heard the premature click.
>>> I ran it using pulseaudio and heard the click.  It seems to be not
>>> consistent to my ear how much of a click, which makes me think there's
>>> thread timing involved where more of the audio may or may not get
>>> thorough before closing.
>>>
>>> In addition, I rendered it to disk and played using the gnome media
>>> player and it did not have a click.
>>>
>>> Note, the build is a float build. (My tests on WinXP are with a
>>> doubles build).
>>>
>>> Going to test on Ubuntu now.
>>>
>>> steven
>>>
>>> On Wed, Jan 20, 2010 at 5:33 PM, Steven Yi  wrote:
>>>> Well, even if ksmps is high, it should still render in realtime
>>>> completely through the event, and may be a clue.  I'm going to test
>>>> on
>>>> fedora and ubuntu virtual machines.  I think Fedora has csound 5.10
>>>> and will use whatever is in Ubuntu (i assume 5.11.1 from Rene's
>>>> setup). I'll use the same test.csd I originally posted. Will report
>>>> back when I have results.
>>>>
>>>> On Wed, Jan 20, 2010 at 5:16 PM, Victor Lazzarini
>>>>  wrote:
>>>>> I think rendering means 'to disk'. His CSD had a very large ksmps.
>>>>> There is no other CSD example with normal ksmps.
>>>>> I understood 'the problem when ksmps is high' in 'realtime and
>>>>> rendering [to disk]' is what he meant.
>>>>>
>>>>> Victor
>>>>>
>>>>> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>>>>>
>>>>>> I just reread Rene's email and he mentioned:
>>>>>>
>>>>>> "I can see the problem in real time and also in rendering when
>>>>>> ksmps
>>>>>> is high."
>>>>>>
>>>>>> I did not see any mention of rendering to disk anywhere.
>>>>>>
>>>>>> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>>>>>>  wrote:
>>>>>>> No. I think what he reported has nothing to do with this. His test
>>>>>>> was
>>>>>>> with high ksmps and the problem was rendering to disk. I think the
>>>>>>> behaviour he reports is the expected. What you have reported is
>>>>>>> with
>>>>>>> normal conditions of ksmps and shows a difference between
>>>>>>> rendering
>>>>>>> to
>>>>>>> disk and RT.
>>>>>>>
>>>>>>> There are no reports this was happening on 5.11. That does not
>>>>>>> mean
>>>>>>> it
>>>>>>> was not.
>>>>>>>
>>>>>>> Victor
>>>>>>>
>>>>>>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>>>>>>
>>>>>>>> Hi John,
>>>>>>>>
>>>>>>>> Rene reported the issue in 5.11.1 for both float and double on
>>>>>>>> Ubuntu
>>>>>>>> Studio 9.10.
>>>>>>>>
>>>>>>>> steven
>>>>>>>>
>>>>>>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>>>>>>> Was this in 5/.11 or has something shaken?
>>>>>>>>>
>>>>>>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a
>>>>>>>>>> sleep(1)
>>>>>>>>>> before stopping the audio callback.
>>>>>>>>>> In the case of pa_cb, it could be the same problem. The
>>>>>>>>>> callback
>>>>>>>>>> is
>>>>>>>>>> being stopped before it's completely done.
>>>>>>>>>> Same on linux(?) I guess this is more apparent with large
>>>>>>>>>> buffers.
>>>>>>>>>>
>>>>>>>>>> Victor
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>> attracts the
>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>> opportunities for
>>>>>>>>> Conference
>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>> issues through
>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>> companies.
>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>> _______________________________________________
>>>>>>>>> Csound-devel mailing list
>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>> attracts
>>>>>>>> the
>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>> for
>>>>>>>> Conference
>>>>>>>> attendees to learn about information security's most important
>>>>>>>> issues through
>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>> companies.
>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>> _______________________________________________
>>>>>>>> Csound-devel mailing list
>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts the
>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>> for
>>>>>>> Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts
>>>>>> the
>>>>>> world's best and brightest in the field, creating opportunities for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities
>>>>> for Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts
>>> the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts the
>> world's best and brightest in the field, creating opportunities for Conference
>> attendees to learn about information security's most important issues through
>> interactions with peers, luminaries and emerging and established companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Sorry for the false alarm, I see that csoundReset is called from
csoundDestroy.  (Sidenote: The csound.h header probably needs updating
as there is a comment there that csoundReset should be called
directly). Looking into csoundReset now.

On Wed, Jan 20, 2010 at 7:50 PM, Steven Yi  wrote:
> Okay, I'm seeing that csound_main.c calls csoundDestroy but does not
> call csoundCleanup. I'm pretty sure it is supposed to: can anyone
> confirm?  Going go test this now.
>
> On Wed, Jan 20, 2010 at 7:29 PM, Steven Yi  wrote:
>> Well, from the looks of it, it seems a problem that is existing in
>> multiple drivers.  I'm wondering if the problem lies elsewhere and is
>> a flaw in Csound that if fixed would fix all the issue at once.  It
>> seems to go back to at least 5.10 from what I found on Ubuntu.  I want
>> to understand the interaction between Csound and the audio output a
>> bit more as I'd like to know exactly why this is happening. I'm
>> worried if we just put in a sleep that it may mask the problem and
>> that it may come up again elsewhere if we don't address the core
>> issue.  (Not to say that we may just need to do a sleep in the end,
>> but at the moment I just don't know what the root cause is.)
>>
>> On Wed, Jan 20, 2010 at 6:44 PM, Victor Lazzarini
>>  wrote:
>>> Why don't  you try to stick  a sleep() or similar  call before any
>>> call to
>>> Pa_StopStream(dev->handle) just to check if that changes things (as it
>>> does in rtcoreaudio())?
>>>
>>> Victor
>>> On 20 Jan 2010, at 23:17, Steven Yi wrote:
>>>
>>>> Okay I just tested using:
>>>>
>>>> Csound version 5.10 beta (float samples) Dec  3 2009
>>>>
>>>> This is on Fedora 12, using what is in the RPM repository which is:
>>>>
>>>> csound-5.10.1-14.fc12.i686
>>>>
>>>> I tested with alsa using -b1024 -B4096 and heard the premature click.
>>>> I ran it using pulseaudio and heard the click.  It seems to be not
>>>> consistent to my ear how much of a click, which makes me think there's
>>>> thread timing involved where more of the audio may or may not get
>>>> thorough before closing.
>>>>
>>>> In addition, I rendered it to disk and played using the gnome media
>>>> player and it did not have a click.
>>>>
>>>> Note, the build is a float build. (My tests on WinXP are with a
>>>> doubles build).
>>>>
>>>> Going to test on Ubuntu now.
>>>>
>>>> steven
>>>>
>>>> On Wed, Jan 20, 2010 at 5:33 PM, Steven Yi  wrote:
>>>>> Well, even if ksmps is high, it should still render in realtime
>>>>> completely through the event, and may be a clue.  I'm going to test
>>>>> on
>>>>> fedora and ubuntu virtual machines.  I think Fedora has csound 5.10
>>>>> and will use whatever is in Ubuntu (i assume 5.11.1 from Rene's
>>>>> setup). I'll use the same test.csd I originally posted. Will report
>>>>> back when I have results.
>>>>>
>>>>> On Wed, Jan 20, 2010 at 5:16 PM, Victor Lazzarini
>>>>>  wrote:
>>>>>> I think rendering means 'to disk'. His CSD had a very large ksmps.
>>>>>> There is no other CSD example with normal ksmps.
>>>>>> I understood 'the problem when ksmps is high' in 'realtime and
>>>>>> rendering [to disk]' is what he meant.
>>>>>>
>>>>>> Victor
>>>>>>
>>>>>> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>>>>>>
>>>>>>> I just reread Rene's email and he mentioned:
>>>>>>>
>>>>>>> "I can see the problem in real time and also in rendering when
>>>>>>> ksmps
>>>>>>> is high."
>>>>>>>
>>>>>>> I did not see any mention of rendering to disk anywhere.
>>>>>>>
>>>>>>> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>>>>>>>  wrote:
>>>>>>>> No. I think what he reported has nothing to do with this. His test
>>>>>>>> was
>>>>>>>> with high ksmps and the problem was rendering to disk. I think the
>>>>>>>> behaviour he reports is the expected. What you have reported is
>>>>>>>> with
>>>>>>>> normal conditions of ksmps and shows a difference between
>>>>>>>> rendering
>>>>>>>> to
>>>>>>>> disk and RT.
>>>>>>>>
>>>>>>>> There are no reports this was happening on 5.11. That does not
>>>>>>>> mean
>>>>>>>> it
>>>>>>>> was not.
>>>>>>>>
>>>>>>>> Victor
>>>>>>>>
>>>>>>>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>>>>>>>
>>>>>>>>> Hi John,
>>>>>>>>>
>>>>>>>>> Rene reported the issue in 5.11.1 for both float and double on
>>>>>>>>> Ubuntu
>>>>>>>>> Studio 9.10.
>>>>>>>>>
>>>>>>>>> steven
>>>>>>>>>
>>>>>>>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>>>>>>>> Was this in 5/.11 or has something shaken?
>>>>>>>>>>
>>>>>>>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a
>>>>>>>>>>> sleep(1)
>>>>>>>>>>> before stopping the audio callback.
>>>>>>>>>>> In the case of pa_cb, it could be the same problem. The
>>>>>>>>>>> callback
>>>>>>>>>>> is
>>>>>>>>>>> being stopped before it's completely done.
>>>>>>>>>>> Same on linux(?) I guess this is more apparent with large
>>>>>>>>>>> buffers.
>>>>>>>>>>>
>>>>>>>>>>> Victor
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>> attracts the
>>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>>> opportunities for
>>>>>>>>>> Conference
>>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>>> issues through
>>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>>> companies.
>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>> attracts
>>>>>>>>> the
>>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>>> for
>>>>>>>>> Conference
>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>> issues through
>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>> companies.
>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>> _______________________________________________
>>>>>>>>> Csound-devel mailing list
>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>> attracts the
>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>> for
>>>>>>>> Conference
>>>>>>>> attendees to learn about information security's most important
>>>>>>>> issues through
>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>> companies.
>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>> _______________________________________________
>>>>>>>> Csound-devel mailing list
>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts
>>>>>>> the
>>>>>>> world's best and brightest in the field, creating opportunities for
>>>>>>> Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts the
>>>>>> world's best and brightest in the field, creating opportunities
>>>>>> for Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently attracts
>>>> the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts the
>>> world's best and brightest in the field, creating opportunities for Conference
>>> attendees to learn about information security's most important issues through
>>> interactions with peers, luminaries and emerging and established companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hi All,

I'm looking at libsnd.c, particularly sfcloseout.  It has this code:

    if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT)) > 0) {
      /* flush outbuffer */
      csound->nrecs++;
      csound->audtran(csound, ST(outbuf), nb);
    }

On this machine I had O->outbufsamps = 8192, and using this before the
statement block above:

    csound->Message(csound, ">>>> %d", ST(outbufrem));

I got an outbufrem of 136854.    The code looks like this all the way
back the first revision of libsnd.c in CVS but at that time outbufrem
was a static variable and I assume was used elsewhere (in that version
of the file, it is only used in this block of code, not set anywhere).
 I am wondering if something here changed causing an issue?  I also
saw that in the oldest code, it was:

    if ((nb = (O.outbufsamps-outbufrem) *
         O.outsampsiz) > 0)/* flush outbuffer */

using sample size, but it seemed to have changed to MYFLT at some point.

steven


On Wed, Jan 20, 2010 at 8:00 PM, Steven Yi  wrote:
> Sorry for the false alarm, I see that csoundReset is called from
> csoundDestroy.  (Sidenote: The csound.h header probably needs updating
> as there is a comment there that csoundReset should be called
> directly). Looking into csoundReset now.
>
> On Wed, Jan 20, 2010 at 7:50 PM, Steven Yi  wrote:
>> Okay, I'm seeing that csound_main.c calls csoundDestroy but does not
>> call csoundCleanup. I'm pretty sure it is supposed to: can anyone
>> confirm?  Going go test this now.
>>
>> On Wed, Jan 20, 2010 at 7:29 PM, Steven Yi  wrote:
>>> Well, from the looks of it, it seems a problem that is existing in
>>> multiple drivers.  I'm wondering if the problem lies elsewhere and is
>>> a flaw in Csound that if fixed would fix all the issue at once.  It
>>> seems to go back to at least 5.10 from what I found on Ubuntu.  I want
>>> to understand the interaction between Csound and the audio output a
>>> bit more as I'd like to know exactly why this is happening. I'm
>>> worried if we just put in a sleep that it may mask the problem and
>>> that it may come up again elsewhere if we don't address the core
>>> issue.  (Not to say that we may just need to do a sleep in the end,
>>> but at the moment I just don't know what the root cause is.)
>>>
>>> On Wed, Jan 20, 2010 at 6:44 PM, Victor Lazzarini
>>>  wrote:
>>>> Why don't  you try to stick  a sleep() or similar  call before any
>>>> call to
>>>> Pa_StopStream(dev->handle) just to check if that changes things (as it
>>>> does in rtcoreaudio())?
>>>>
>>>> Victor
>>>> On 20 Jan 2010, at 23:17, Steven Yi wrote:
>>>>
>>>>> Okay I just tested using:
>>>>>
>>>>> Csound version 5.10 beta (float samples) Dec  3 2009
>>>>>
>>>>> This is on Fedora 12, using what is in the RPM repository which is:
>>>>>
>>>>> csound-5.10.1-14.fc12.i686
>>>>>
>>>>> I tested with alsa using -b1024 -B4096 and heard the premature click.
>>>>> I ran it using pulseaudio and heard the click.  It seems to be not
>>>>> consistent to my ear how much of a click, which makes me think there's
>>>>> thread timing involved where more of the audio may or may not get
>>>>> thorough before closing.
>>>>>
>>>>> In addition, I rendered it to disk and played using the gnome media
>>>>> player and it did not have a click.
>>>>>
>>>>> Note, the build is a float build. (My tests on WinXP are with a
>>>>> doubles build).
>>>>>
>>>>> Going to test on Ubuntu now.
>>>>>
>>>>> steven
>>>>>
>>>>> On Wed, Jan 20, 2010 at 5:33 PM, Steven Yi  wrote:
>>>>>> Well, even if ksmps is high, it should still render in realtime
>>>>>> completely through the event, and may be a clue.  I'm going to test
>>>>>> on
>>>>>> fedora and ubuntu virtual machines.  I think Fedora has csound 5.10
>>>>>> and will use whatever is in Ubuntu (i assume 5.11.1 from Rene's
>>>>>> setup). I'll use the same test.csd I originally posted. Will report
>>>>>> back when I have results.
>>>>>>
>>>>>> On Wed, Jan 20, 2010 at 5:16 PM, Victor Lazzarini
>>>>>>  wrote:
>>>>>>> I think rendering means 'to disk'. His CSD had a very large ksmps.
>>>>>>> There is no other CSD example with normal ksmps.
>>>>>>> I understood 'the problem when ksmps is high' in 'realtime and
>>>>>>> rendering [to disk]' is what he meant.
>>>>>>>
>>>>>>> Victor
>>>>>>>
>>>>>>> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>>>>>>>
>>>>>>>> I just reread Rene's email and he mentioned:
>>>>>>>>
>>>>>>>> "I can see the problem in real time and also in rendering when
>>>>>>>> ksmps
>>>>>>>> is high."
>>>>>>>>
>>>>>>>> I did not see any mention of rendering to disk anywhere.
>>>>>>>>
>>>>>>>> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>>>>>>>>  wrote:
>>>>>>>>> No. I think what he reported has nothing to do with this. His test
>>>>>>>>> was
>>>>>>>>> with high ksmps and the problem was rendering to disk. I think the
>>>>>>>>> behaviour he reports is the expected. What you have reported is
>>>>>>>>> with
>>>>>>>>> normal conditions of ksmps and shows a difference between
>>>>>>>>> rendering
>>>>>>>>> to
>>>>>>>>> disk and RT.
>>>>>>>>>
>>>>>>>>> There are no reports this was happening on 5.11. That does not
>>>>>>>>> mean
>>>>>>>>> it
>>>>>>>>> was not.
>>>>>>>>>
>>>>>>>>> Victor
>>>>>>>>>
>>>>>>>>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>>>>>>>>
>>>>>>>>>> Hi John,
>>>>>>>>>>
>>>>>>>>>> Rene reported the issue in 5.11.1 for both float and double on
>>>>>>>>>> Ubuntu
>>>>>>>>>> Studio 9.10.
>>>>>>>>>>
>>>>>>>>>> steven
>>>>>>>>>>
>>>>>>>>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>>>>>>>>> Was this in 5/.11 or has something shaken?
>>>>>>>>>>>
>>>>>>>>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a
>>>>>>>>>>>> sleep(1)
>>>>>>>>>>>> before stopping the audio callback.
>>>>>>>>>>>> In the case of pa_cb, it could be the same problem. The
>>>>>>>>>>>> callback
>>>>>>>>>>>> is
>>>>>>>>>>>> being stopped before it's completely done.
>>>>>>>>>>>> Same on linux(?) I guess this is more apparent with large
>>>>>>>>>>>> buffers.
>>>>>>>>>>>>
>>>>>>>>>>>> Victor
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>>> attracts the
>>>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>>>> opportunities for
>>>>>>>>>>> Conference
>>>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>>>> issues through
>>>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>>>> companies.
>>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>> attracts
>>>>>>>>>> the
>>>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>>>> for
>>>>>>>>>> Conference
>>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>>> issues through
>>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>>> companies.
>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>> attracts the
>>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>>> for
>>>>>>>>> Conference
>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>> issues through
>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>> companies.
>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>> _______________________________________________
>>>>>>>>> Csound-devel mailing list
>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>> attracts
>>>>>>>> the
>>>>>>>> world's best and brightest in the field, creating opportunities for
>>>>>>>> Conference
>>>>>>>> attendees to learn about information security's most important
>>>>>>>> issues through
>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>> companies.
>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>> _______________________________________________
>>>>>>>> Csound-devel mailing list
>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts the
>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>> for Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently attracts
>>>>> the
>>>>> world's best and brightest in the field, creating opportunities for
>>>>> Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently attracts the
>>>> world's best and brightest in the field, creating opportunities for Conference
>>>> attendees to learn about information security's most important issues through
>>>> interactions with peers, luminaries and emerging and established companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>
>>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

The only change I can think of was the change from counting in seconds
(floating point) to counting in samples  There could be a > or >= prpoblem
there.  Mainly in musmon.c
  That does not seem to be enough to explain i tthough

==John ff

> Hi All,
>
> I'm looking at libsnd.c, particularly sfcloseout.  It has this code:
>
>     if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT)) > 0) {
>       /* flush outbuffer */
>       csound->nrecs++;
>       csound->audtran(csound, ST(outbuf), nb);
>     }
>
> On this machine I had O->outbufsamps = 8192, and using this before the
> statement block above:
>
>     csound->Message(csound, ">>>> %d", ST(outbufrem));
>
> I got an outbufrem of 136854.    The code looks like this all the way
> back the first revision of libsnd.c in CVS but at that time outbufrem
> was a static variable and I assume was used elsewhere (in that version
> of the file, it is only used in this block of code, not set anywhere).
>  I am wondering if something here changed causing an issue?  I also
> saw that in the oldest code, it was:
>
>     if ((nb = (O.outbufsamps-outbufrem) *
>          O.outsampsiz) > 0)/* flush outbuffer */
>
> using sample size, but it seemed to have changed to MYFLT at some point.
>



------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hmm... not sure if this is a red herring for me.  I rendered to disk
and outbufrem was again 136854.

On Wed, Jan 20, 2010 at 8:46 PM, Steven Yi  wrote:
> Hi All,
>
> I'm looking at libsnd.c, particularly sfcloseout.  It has this code:
>
>    if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT)) > 0) {
>      /* flush outbuffer */
>      csound->nrecs++;
>      csound->audtran(csound, ST(outbuf), nb);
>    }
>
> On this machine I had O->outbufsamps = 8192, and using this before the
> statement block above:
>
>    csound->Message(csound, ">>>> %d", ST(outbufrem));
>
> I got an outbufrem of 136854.    The code looks like this all the way
> back the first revision of libsnd.c in CVS but at that time outbufrem
> was a static variable and I assume was used elsewhere (in that version
> of the file, it is only used in this block of code, not set anywhere).
>  I am wondering if something here changed causing an issue?  I also
> saw that in the oldest code, it was:
>
>    if ((nb = (O.outbufsamps-outbufrem) *
>         O.outsampsiz) > 0)/* flush outbuffer */
>
> using sample size, but it seemed to have changed to MYFLT at some point.
>
> steven
>
>
> On Wed, Jan 20, 2010 at 8:00 PM, Steven Yi  wrote:
>> Sorry for the false alarm, I see that csoundReset is called from
>> csoundDestroy.  (Sidenote: The csound.h header probably needs updating
>> as there is a comment there that csoundReset should be called
>> directly). Looking into csoundReset now.
>>
>> On Wed, Jan 20, 2010 at 7:50 PM, Steven Yi  wrote:
>>> Okay, I'm seeing that csound_main.c calls csoundDestroy but does not
>>> call csoundCleanup. I'm pretty sure it is supposed to: can anyone
>>> confirm?  Going go test this now.
>>>
>>> On Wed, Jan 20, 2010 at 7:29 PM, Steven Yi  wrote:
>>>> Well, from the looks of it, it seems a problem that is existing in
>>>> multiple drivers.  I'm wondering if the problem lies elsewhere and is
>>>> a flaw in Csound that if fixed would fix all the issue at once.  It
>>>> seems to go back to at least 5.10 from what I found on Ubuntu.  I want
>>>> to understand the interaction between Csound and the audio output a
>>>> bit more as I'd like to know exactly why this is happening. I'm
>>>> worried if we just put in a sleep that it may mask the problem and
>>>> that it may come up again elsewhere if we don't address the core
>>>> issue.  (Not to say that we may just need to do a sleep in the end,
>>>> but at the moment I just don't know what the root cause is.)
>>>>
>>>> On Wed, Jan 20, 2010 at 6:44 PM, Victor Lazzarini
>>>>  wrote:
>>>>> Why don't  you try to stick  a sleep() or similar  call before any
>>>>> call to
>>>>> Pa_StopStream(dev->handle) just to check if that changes things (as it
>>>>> does in rtcoreaudio())?
>>>>>
>>>>> Victor
>>>>> On 20 Jan 2010, at 23:17, Steven Yi wrote:
>>>>>
>>>>>> Okay I just tested using:
>>>>>>
>>>>>> Csound version 5.10 beta (float samples) Dec  3 2009
>>>>>>
>>>>>> This is on Fedora 12, using what is in the RPM repository which is:
>>>>>>
>>>>>> csound-5.10.1-14.fc12.i686
>>>>>>
>>>>>> I tested with alsa using -b1024 -B4096 and heard the premature click.
>>>>>> I ran it using pulseaudio and heard the click.  It seems to be not
>>>>>> consistent to my ear how much of a click, which makes me think there's
>>>>>> thread timing involved where more of the audio may or may not get
>>>>>> thorough before closing.
>>>>>>
>>>>>> In addition, I rendered it to disk and played using the gnome media
>>>>>> player and it did not have a click.
>>>>>>
>>>>>> Note, the build is a float build. (My tests on WinXP are with a
>>>>>> doubles build).
>>>>>>
>>>>>> Going to test on Ubuntu now.
>>>>>>
>>>>>> steven
>>>>>>
>>>>>> On Wed, Jan 20, 2010 at 5:33 PM, Steven Yi  wrote:
>>>>>>> Well, even if ksmps is high, it should still render in realtime
>>>>>>> completely through the event, and may be a clue.  I'm going to test
>>>>>>> on
>>>>>>> fedora and ubuntu virtual machines.  I think Fedora has csound 5.10
>>>>>>> and will use whatever is in Ubuntu (i assume 5.11.1 from Rene's
>>>>>>> setup). I'll use the same test.csd I originally posted. Will report
>>>>>>> back when I have results.
>>>>>>>
>>>>>>> On Wed, Jan 20, 2010 at 5:16 PM, Victor Lazzarini
>>>>>>>  wrote:
>>>>>>>> I think rendering means 'to disk'. His CSD had a very large ksmps.
>>>>>>>> There is no other CSD example with normal ksmps.
>>>>>>>> I understood 'the problem when ksmps is high' in 'realtime and
>>>>>>>> rendering [to disk]' is what he meant.
>>>>>>>>
>>>>>>>> Victor
>>>>>>>>
>>>>>>>> On 20 Jan 2010, at 22:10, Steven Yi wrote:
>>>>>>>>
>>>>>>>>> I just reread Rene's email and he mentioned:
>>>>>>>>>
>>>>>>>>> "I can see the problem in real time and also in rendering when
>>>>>>>>> ksmps
>>>>>>>>> is high."
>>>>>>>>>
>>>>>>>>> I did not see any mention of rendering to disk anywhere.
>>>>>>>>>
>>>>>>>>> On Wed, Jan 20, 2010 at 4:57 PM, Victor Lazzarini
>>>>>>>>>  wrote:
>>>>>>>>>> No. I think what he reported has nothing to do with this. His test
>>>>>>>>>> was
>>>>>>>>>> with high ksmps and the problem was rendering to disk. I think the
>>>>>>>>>> behaviour he reports is the expected. What you have reported is
>>>>>>>>>> with
>>>>>>>>>> normal conditions of ksmps and shows a difference between
>>>>>>>>>> rendering
>>>>>>>>>> to
>>>>>>>>>> disk and RT.
>>>>>>>>>>
>>>>>>>>>> There are no reports this was happening on 5.11. That does not
>>>>>>>>>> mean
>>>>>>>>>> it
>>>>>>>>>> was not.
>>>>>>>>>>
>>>>>>>>>> Victor
>>>>>>>>>>
>>>>>>>>>> On 20 Jan 2010, at 21:51, Steven Yi wrote:
>>>>>>>>>>
>>>>>>>>>>> Hi John,
>>>>>>>>>>>
>>>>>>>>>>> Rene reported the issue in 5.11.1 for both float and double on
>>>>>>>>>>> Ubuntu
>>>>>>>>>>> Studio 9.10.
>>>>>>>>>>>
>>>>>>>>>>> steven
>>>>>>>>>>>
>>>>>>>>>>> On Wed, Jan 20, 2010 at 4:49 PM,   wrote:
>>>>>>>>>>>> Was this in 5/.11 or has something shaken?
>>>>>>>>>>>>
>>>>>>>>>>>>> It seems to be fixed (on rtcoreaudio.c, if I just place a
>>>>>>>>>>>>> sleep(1)
>>>>>>>>>>>>> before stopping the audio callback.
>>>>>>>>>>>>> In the case of pa_cb, it could be the same problem. The
>>>>>>>>>>>>> callback
>>>>>>>>>>>>> is
>>>>>>>>>>>>> being stopped before it's completely done.
>>>>>>>>>>>>> Same on linux(?) I guess this is more apparent with large
>>>>>>>>>>>>> buffers.
>>>>>>>>>>>>>
>>>>>>>>>>>>> Victor
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>>>> attracts the
>>>>>>>>>>>> world's best and brightest in the field, creating
>>>>>>>>>>>> opportunities for
>>>>>>>>>>>> Conference
>>>>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>>>>> issues through
>>>>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>>>>> companies.
>>>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>>>>> _______________________________________________
>>>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>>>>
>>>>>>>>>>>
>>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>>> attracts
>>>>>>>>>>> the
>>>>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>>>>> for
>>>>>>>>>>> Conference
>>>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>>>> issues through
>>>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>>>> companies.
>>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>>>> _______________________________________________
>>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>>> attracts the
>>>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>>>> for
>>>>>>>>>> Conference
>>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>>> issues through
>>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>>> companies.
>>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>>> _______________________________________________
>>>>>>>>>> Csound-devel mailing list
>>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>>> attracts
>>>>>>>>> the
>>>>>>>>> world's best and brightest in the field, creating opportunities for
>>>>>>>>> Conference
>>>>>>>>> attendees to learn about information security's most important
>>>>>>>>> issues through
>>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>>> companies.
>>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>>> _______________________________________________
>>>>>>>>> Csound-devel mailing list
>>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>
>>>>>>>>
>>>>>>>> ------------------------------------------------------------------------------
>>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>>> attracts the
>>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>>> for Conference
>>>>>>>> attendees to learn about information security's most important
>>>>>>>> issues through
>>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>>> companies.
>>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>>> _______________________________________________
>>>>>>>> Csound-devel mailing list
>>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently attracts
>>>>>> the
>>>>>> world's best and brightest in the field, creating opportunities for
>>>>>> Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently attracts the
>>>>> world's best and brightest in the field, creating opportunities for Conference
>>>>> attendees to learn about information security's most important issues through
>>>>> interactions with peers, luminaries and emerging and established companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>
>>>
>>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Yeah, I'm thinking now though if that was a problem then it'd affect
output to disk.  I should have thought earlier to remember that since
one of the audio drivers is working here (pa_bl) then it would mean
csound's code should be fine. I'm beginning to think that we just need
to go through each audio driver and make sure it is flushing properly
before closing.  I was experimenting with modifications to rtwinmm.c's
rtclose and managed to hear more of the output though still slightly
chopped off.  I also noticed the pulseaudio driver does not do a flush
with pa_simple_drain() before close (going to try this out now on
Fedora on my netbook).


On Wed, Jan 20, 2010 at 9:48 PM,   wrote:
> The only change I can think of was the change from counting in seconds
> (floating point) to counting in samples  There could be a > or >= prpoblem
> there.  Mainly in musmon.c
>  That does not seem to be enough to explain i tthough
>
> ==John ff
>
>> Hi All,
>>
>> I'm looking at libsnd.c, particularly sfcloseout.  It has this code:
>>
>>     if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT)) > 0) {
>>       /* flush outbuffer */
>>       csound->nrecs++;
>>       csound->audtran(csound, ST(outbuf), nb);
>>     }
>>
>> On this machine I had O->outbufsamps = 8192, and using this before the
>> statement block above:
>>
>>     csound->Message(csound, ">>>> %d", ST(outbufrem));
>>
>> I got an outbufrem of 136854.    The code looks like this all the way
>> back the first revision of libsnd.c in CVS but at that time outbufrem
>> was a static variable and I assume was used elsewhere (in that version
>> of the file, it is only used in this block of code, not set anywhere).
>>  I am wondering if something here changed causing an issue?  I also
>> saw that in the oldest code, it was:
>>
>>     if ((nb = (O.outbufsamps-outbufrem) *
>>          O.outsampsiz) > 0)/* flush outbuffer */
>>
>> using sample size, but it seemed to have changed to MYFLT at some point.
>>
>
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Adding the flush to pulseaudio fixed the issue for that driver
(committed to CVS).  I am guessing at this point it is a problem that
is in different drivers and we'll have to go and check each.  Seems
like we should just go through all of the drivers and see if the audio
API each driver is using has some means of draining or flushing out
the buffer before closing, and if we're doing it correctly.

On Wed, Jan 20, 2010 at 10:43 PM, Steven Yi  wrote:
> Yeah, I'm thinking now though if that was a problem then it'd affect
> output to disk.  I should have thought earlier to remember that since
> one of the audio drivers is working here (pa_bl) then it would mean
> csound's code should be fine. I'm beginning to think that we just need
> to go through each audio driver and make sure it is flushing properly
> before closing.  I was experimenting with modifications to rtwinmm.c's
> rtclose and managed to hear more of the output though still slightly
> chopped off.  I also noticed the pulseaudio driver does not do a flush
> with pa_simple_drain() before close (going to try this out now on
> Fedora on my netbook).
>
>
> On Wed, Jan 20, 2010 at 9:48 PM,   wrote:
>> The only change I can think of was the change from counting in seconds
>> (floating point) to counting in samples  There could be a > or >= prpoblem
>> there.  Mainly in musmon.c
>>  That does not seem to be enough to explain i tthough
>>
>> ==John ff
>>
>>> Hi All,
>>>
>>> I'm looking at libsnd.c, particularly sfcloseout.  It has this code:
>>>
>>>     if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT)) > 0) {
>>>       /* flush outbuffer */
>>>       csound->nrecs++;
>>>       csound->audtran(csound, ST(outbuf), nb);
>>>     }
>>>
>>> On this machine I had O->outbufsamps = 8192, and using this before the
>>> statement block above:
>>>
>>>     csound->Message(csound, ">>>> %d", ST(outbufrem));
>>>
>>> I got an outbufrem of 136854.    The code looks like this all the way
>>> back the first revision of libsnd.c in CVS but at that time outbufrem
>>> was a static variable and I assume was used elsewhere (in that version
>>> of the file, it is only used in this block of code, not set anywhere).
>>>  I am wondering if something here changed causing an issue?  I also
>>> saw that in the oldest code, it was:
>>>
>>>     if ((nb = (O.outbufsamps-outbufrem) *
>>>          O.outsampsiz) > 0)/* flush outbuffer */
>>>
>>> using sample size, but it seemed to have changed to MYFLT at some point.
>>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts the
>> world's best and brightest in the field, creating opportunities for Conference
>> attendees to learn about information security's most important issues through
>> interactions with peers, luminaries and emerging and established companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hi All,

I've fiddled with the portaudio driver callback closing this morning
and haven't managed to find a way to get it properly wait to close.
Strangely enough, I've added Pa_Sleep all around the closing code in
rtclose_ and it doesn't seem to remove the pop.  I've tried all manner
of changes.  I'm going to give up on this for now and move to see
about removing pops in the other drivers.

Thanks!
steven

On Wed, Jan 20, 2010 at 11:46 PM, Steven Yi  wrote:
> Adding the flush to pulseaudio fixed the issue for that driver
> (committed to CVS).  I am guessing at this point it is a problem that
> is in different drivers and we'll have to go and check each.  Seems
> like we should just go through all of the drivers and see if the audio
> API each driver is using has some means of draining or flushing out
> the buffer before closing, and if we're doing it correctly.
>
> On Wed, Jan 20, 2010 at 10:43 PM, Steven Yi  wrote:
>> Yeah, I'm thinking now though if that was a problem then it'd affect
>> output to disk.  I should have thought earlier to remember that since
>> one of the audio drivers is working here (pa_bl) then it would mean
>> csound's code should be fine. I'm beginning to think that we just need
>> to go through each audio driver and make sure it is flushing properly
>> before closing.  I was experimenting with modifications to rtwinmm.c's
>> rtclose and managed to hear more of the output though still slightly
>> chopped off.  I also noticed the pulseaudio driver does not do a flush
>> with pa_simple_drain() before close (going to try this out now on
>> Fedora on my netbook).
>>
>>
>> On Wed, Jan 20, 2010 at 9:48 PM,   wrote:
>>> The only change I can think of was the change from counting in seconds
>>> (floating point) to counting in samples  There could be a > or >= prpoblem
>>> there.  Mainly in musmon.c
>>>  That does not seem to be enough to explain i tthough
>>>
>>> ==John ff
>>>
>>>> Hi All,
>>>>
>>>> I'm looking at libsnd.c, particularly sfcloseout.  It has this code:
>>>>
>>>>     if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT)) > 0) {
>>>>       /* flush outbuffer */
>>>>       csound->nrecs++;
>>>>       csound->audtran(csound, ST(outbuf), nb);
>>>>     }
>>>>
>>>> On this machine I had O->outbufsamps = 8192, and using this before the
>>>> statement block above:
>>>>
>>>>     csound->Message(csound, ">>>> %d", ST(outbufrem));
>>>>
>>>> I got an outbufrem of 136854.    The code looks like this all the way
>>>> back the first revision of libsnd.c in CVS but at that time outbufrem
>>>> was a static variable and I assume was used elsewhere (in that version
>>>> of the file, it is only used in this block of code, not set anywhere).
>>>>  I am wondering if something here changed causing an issue?  I also
>>>> saw that in the oldest code, it was:
>>>>
>>>>     if ((nb = (O.outbufsamps-outbufrem) *
>>>>          O.outsampsiz) > 0)/* flush outbuffer */
>>>>
>>>> using sample size, but it seemed to have changed to MYFLT at some point.
>>>>
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts the
>>> world's best and brightest in the field, creating opportunities for Conference
>>> attendees to learn about information security's most important issues through
>>> interactions with peers, luminaries and emerging and established companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Really strange that it does not happen on OSX. It is probably the  
underlying API (mme, alsa)
behind portaudio. Does it happen with ASIO or just with MME?

Victor

On 21 Jan 2010, at 17:40, Steven Yi wrote:

> Hi All,
>
> I've fiddled with the portaudio driver callback closing this morning
> and haven't managed to find a way to get it properly wait to close.
> Strangely enough, I've added Pa_Sleep all around the closing code in
> rtclose_ and it doesn't seem to remove the pop.  I've tried all manner
> of changes.  I'm going to give up on this for now and move to see
> about removing pops in the other drivers.
>
> Thanks!
> steven
>
> On Wed, Jan 20, 2010 at 11:46 PM, Steven Yi   
> wrote:
>> Adding the flush to pulseaudio fixed the issue for that driver
>> (committed to CVS).  I am guessing at this point it is a problem that
>> is in different drivers and we'll have to go and check each.  Seems
>> like we should just go through all of the drivers and see if the  
>> audio
>> API each driver is using has some means of draining or flushing out
>> the buffer before closing, and if we're doing it correctly.
>>
>> On Wed, Jan 20, 2010 at 10:43 PM, Steven Yi   
>> wrote:
>>> Yeah, I'm thinking now though if that was a problem then it'd affect
>>> output to disk.  I should have thought earlier to remember that  
>>> since
>>> one of the audio drivers is working here (pa_bl) then it would mean
>>> csound's code should be fine. I'm beginning to think that we just  
>>> need
>>> to go through each audio driver and make sure it is flushing  
>>> properly
>>> before closing.  I was experimenting with modifications to  
>>> rtwinmm.c's
>>> rtclose and managed to hear more of the output though still slightly
>>> chopped off.  I also noticed the pulseaudio driver does not do a  
>>> flush
>>> with pa_simple_drain() before close (going to try this out now on
>>> Fedora on my netbook).
>>>
>>>
>>> On Wed, Jan 20, 2010 at 9:48 PM,   wrote:
>>>> The only change I can think of was the change from counting in  
>>>> seconds
>>>> (floating point) to counting in samples  There could be a > or >=  
>>>> prpoblem
>>>> there.  Mainly in musmon.c
>>>>  That does not seem to be enough to explain i tthough
>>>>
>>>> ==John ff
>>>>
>>>>> Hi All,
>>>>>
>>>>> I'm looking at libsnd.c, particularly sfcloseout.  It has this  
>>>>> code:
>>>>>
>>>>>     if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT))  
>>>>> > 0) {
>>>>>       /* flush outbuffer */
>>>>>       csound->nrecs++;
>>>>>       csound->audtran(csound, ST(outbuf), nb);
>>>>>     }
>>>>>
>>>>> On this machine I had O->outbufsamps = 8192, and using this  
>>>>> before the
>>>>> statement block above:
>>>>>
>>>>>     csound->Message(csound, ">>>> %d", ST(outbufrem));
>>>>>
>>>>> I got an outbufrem of 136854.    The code looks like this all  
>>>>> the way
>>>>> back the first revision of libsnd.c in CVS but at that time  
>>>>> outbufrem
>>>>> was a static variable and I assume was used elsewhere (in that  
>>>>> version
>>>>> of the file, it is only used in this block of code, not set  
>>>>> anywhere).
>>>>>  I am wondering if something here changed causing an issue?  I  
>>>>> also
>>>>> saw that in the oldest code, it was:
>>>>>
>>>>>     if ((nb = (O.outbufsamps-outbufrem) *
>>>>>          O.outsampsiz) > 0)/* flush outbuffer */
>>>>>
>>>>> using sample size, but it seemed to have changed to MYFLT at  
>>>>> some point.
>>>>>
>>>>
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently  
>>>> attracts the
>>>> world's best and brightest in the field, creating opportunities  
>>>> for Conference
>>>> attendees to learn about information security's most important  
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established  
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>
>>>
>>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I only have MME here for PA.  ASIO4ALL causes some weird issue on this
computer.  I think you're right though that it must have to do with
the underlying API.  Interesting to note the blocking API works with
MME.  I may take a look at what is going on in the portaudio blocking
for MME to fix up our MME driver.  I'll try to take a look at Jack
next as Andres reported that being an issue.

On Thu, Jan 21, 2010 at 12:51 PM, Victor Lazzarini
 wrote:
> Really strange that it does not happen on OSX. It is probably the
> underlying API (mme, alsa)
> behind portaudio. Does it happen with ASIO or just with MME?
>
> Victor
>
> On 21 Jan 2010, at 17:40, Steven Yi wrote:
>
>> Hi All,
>>
>> I've fiddled with the portaudio driver callback closing this morning
>> and haven't managed to find a way to get it properly wait to close.
>> Strangely enough, I've added Pa_Sleep all around the closing code in
>> rtclose_ and it doesn't seem to remove the pop.  I've tried all manner
>> of changes.  I'm going to give up on this for now and move to see
>> about removing pops in the other drivers.
>>
>> Thanks!
>> steven
>>
>> On Wed, Jan 20, 2010 at 11:46 PM, Steven Yi 
>> wrote:
>>> Adding the flush to pulseaudio fixed the issue for that driver
>>> (committed to CVS).  I am guessing at this point it is a problem that
>>> is in different drivers and we'll have to go and check each.  Seems
>>> like we should just go through all of the drivers and see if the
>>> audio
>>> API each driver is using has some means of draining or flushing out
>>> the buffer before closing, and if we're doing it correctly.
>>>
>>> On Wed, Jan 20, 2010 at 10:43 PM, Steven Yi 
>>> wrote:
>>>> Yeah, I'm thinking now though if that was a problem then it'd affect
>>>> output to disk.  I should have thought earlier to remember that
>>>> since
>>>> one of the audio drivers is working here (pa_bl) then it would mean
>>>> csound's code should be fine. I'm beginning to think that we just
>>>> need
>>>> to go through each audio driver and make sure it is flushing
>>>> properly
>>>> before closing.  I was experimenting with modifications to
>>>> rtwinmm.c's
>>>> rtclose and managed to hear more of the output though still slightly
>>>> chopped off.  I also noticed the pulseaudio driver does not do a
>>>> flush
>>>> with pa_simple_drain() before close (going to try this out now on
>>>> Fedora on my netbook).
>>>>
>>>>
>>>> On Wed, Jan 20, 2010 at 9:48 PM,   wrote:
>>>>> The only change I can think of was the change from counting in
>>>>> seconds
>>>>> (floating point) to counting in samples  There could be a > or >=
>>>>> prpoblem
>>>>> there.  Mainly in musmon.c
>>>>>  That does not seem to be enough to explain i tthough
>>>>>
>>>>> ==John ff
>>>>>
>>>>>> Hi All,
>>>>>>
>>>>>> I'm looking at libsnd.c, particularly sfcloseout.  It has this
>>>>>> code:
>>>>>>
>>>>>>     if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT))
>>>>>> > 0) {
>>>>>>       /* flush outbuffer */
>>>>>>       csound->nrecs++;
>>>>>>       csound->audtran(csound, ST(outbuf), nb);
>>>>>>     }
>>>>>>
>>>>>> On this machine I had O->outbufsamps = 8192, and using this
>>>>>> before the
>>>>>> statement block above:
>>>>>>
>>>>>>     csound->Message(csound, ">>>> %d", ST(outbufrem));
>>>>>>
>>>>>> I got an outbufrem of 136854.    The code looks like this all
>>>>>> the way
>>>>>> back the first revision of libsnd.c in CVS but at that time
>>>>>> outbufrem
>>>>>> was a static variable and I assume was used elsewhere (in that
>>>>>> version
>>>>>> of the file, it is only used in this block of code, not set
>>>>>> anywhere).
>>>>>>  I am wondering if something here changed causing an issue?  I
>>>>>> also
>>>>>> saw that in the oldest code, it was:
>>>>>>
>>>>>>     if ((nb = (O.outbufsamps-outbufrem) *
>>>>>>          O.outsampsiz) > 0)/* flush outbuffer */
>>>>>>
>>>>>> using sample size, but it seemed to have changed to MYFLT at
>>>>>> some point.
>>>>>>
>>>>>
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------------
>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>> attracts the
>>>>> world's best and brightest in the field, creating opportunities
>>>>> for Conference
>>>>> attendees to learn about information security's most important
>>>>> issues through
>>>>> interactions with peers, luminaries and emerging and established
>>>>> companies.
>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>> _______________________________________________
>>>>> Csound-devel mailing list
>>>>> Csound-devel@lists.sourceforge.net
>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>
>>>>
>>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts
>> the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important
>> issues through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Again, Jack is OK on OSX as I did the same test as Andres and did not  
see anything wrong. It seems OSX escaped most of the issues you  
reported (apart from the coreaudio module, but that is fixed now).

Do you think I can move to build the binaries or should I wait? Are  
the sources going to be retagged?

Victor
On 21 Jan 2010, at 22:09, Steven Yi wrote:

> I only have MME here for PA.  ASIO4ALL causes some weird issue on this
> computer.  I think you're right though that it must have to do with
> the underlying API.  Interesting to note the blocking API works with
> MME.  I may take a look at what is going on in the portaudio blocking
> for MME to fix up our MME driver.  I'll try to take a look at Jack
> next as Andres reported that being an issue.
>
> On Thu, Jan 21, 2010 at 12:51 PM, Victor Lazzarini
>  wrote:
>> Really strange that it does not happen on OSX. It is probably the
>> underlying API (mme, alsa)
>> behind portaudio. Does it happen with ASIO or just with MME?
>>
>> Victor
>>
>> On 21 Jan 2010, at 17:40, Steven Yi wrote:
>>
>>> Hi All,
>>>
>>> I've fiddled with the portaudio driver callback closing this morning
>>> and haven't managed to find a way to get it properly wait to close.
>>> Strangely enough, I've added Pa_Sleep all around the closing code in
>>> rtclose_ and it doesn't seem to remove the pop.  I've tried all  
>>> manner
>>> of changes.  I'm going to give up on this for now and move to see
>>> about removing pops in the other drivers.
>>>
>>> Thanks!
>>> steven
>>>
>>> On Wed, Jan 20, 2010 at 11:46 PM, Steven Yi 
>>> wrote:
>>>> Adding the flush to pulseaudio fixed the issue for that driver
>>>> (committed to CVS).  I am guessing at this point it is a problem  
>>>> that
>>>> is in different drivers and we'll have to go and check each.  Seems
>>>> like we should just go through all of the drivers and see if the
>>>> audio
>>>> API each driver is using has some means of draining or flushing out
>>>> the buffer before closing, and if we're doing it correctly.
>>>>
>>>> On Wed, Jan 20, 2010 at 10:43 PM, Steven Yi 
>>>> wrote:
>>>>> Yeah, I'm thinking now though if that was a problem then it'd  
>>>>> affect
>>>>> output to disk.  I should have thought earlier to remember that
>>>>> since
>>>>> one of the audio drivers is working here (pa_bl) then it would  
>>>>> mean
>>>>> csound's code should be fine. I'm beginning to think that we just
>>>>> need
>>>>> to go through each audio driver and make sure it is flushing
>>>>> properly
>>>>> before closing.  I was experimenting with modifications to
>>>>> rtwinmm.c's
>>>>> rtclose and managed to hear more of the output though still  
>>>>> slightly
>>>>> chopped off.  I also noticed the pulseaudio driver does not do a
>>>>> flush
>>>>> with pa_simple_drain() before close (going to try this out now on
>>>>> Fedora on my netbook).
>>>>>
>>>>>
>>>>> On Wed, Jan 20, 2010 at 9:48 PM,   wrote:
>>>>>> The only change I can think of was the change from counting in
>>>>>> seconds
>>>>>> (floating point) to counting in samples  There could be a > or >=
>>>>>> prpoblem
>>>>>> there.  Mainly in musmon.c
>>>>>>  That does not seem to be enough to explain i tthough
>>>>>>
>>>>>> ==John ff
>>>>>>
>>>>>>> Hi All,
>>>>>>>
>>>>>>> I'm looking at libsnd.c, particularly sfcloseout.  It has this
>>>>>>> code:
>>>>>>>
>>>>>>>     if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT))
>>>>>>>> 0) {
>>>>>>>       /* flush outbuffer */
>>>>>>>       csound->nrecs++;
>>>>>>>       csound->audtran(csound, ST(outbuf), nb);
>>>>>>>     }
>>>>>>>
>>>>>>> On this machine I had O->outbufsamps = 8192, and using this
>>>>>>> before the
>>>>>>> statement block above:
>>>>>>>
>>>>>>>     csound->Message(csound, ">>>> %d", ST(outbufrem));
>>>>>>>
>>>>>>> I got an outbufrem of 136854.    The code looks like this all
>>>>>>> the way
>>>>>>> back the first revision of libsnd.c in CVS but at that time
>>>>>>> outbufrem
>>>>>>> was a static variable and I assume was used elsewhere (in that
>>>>>>> version
>>>>>>> of the file, it is only used in this block of code, not set
>>>>>>> anywhere).
>>>>>>>  I am wondering if something here changed causing an issue?  I
>>>>>>> also
>>>>>>> saw that in the oldest code, it was:
>>>>>>>
>>>>>>>     if ((nb = (O.outbufsamps-outbufrem) *
>>>>>>>          O.outsampsiz) > 0)/* flush outbuffer */
>>>>>>>
>>>>>>> using sample size, but it seemed to have changed to MYFLT at
>>>>>>> some point.
>>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------------
>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>> attracts the
>>>>>> world's best and brightest in the field, creating opportunities
>>>>>> for Conference
>>>>>> attendees to learn about information security's most important
>>>>>> issues through
>>>>>> interactions with peers, luminaries and emerging and established
>>>>>> companies.
>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>> _______________________________________________
>>>>>> Csound-devel mailing list
>>>>>> Csound-devel@lists.sourceforge.net
>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>
>>>>>
>>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently attracts
>>> the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently  
>> attracts the
>> world's best and brightest in the field, creating opportunities for  
>> Conference
>> attendees to learn about information security's most important  
>> issues through
>> interactions with peers, luminaries and emerging and established  
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts  
> the
> world's best and brightest in the field, creating opportunities for  
> Conference
> attendees to learn about information security's most important  
> issues through
> interactions with peers, luminaries and emerging and established  
> companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel


------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hmmm... I just tried my test file with jack on Fedora 12 and did not
get the pops.  If you all want to go ahead with release, that's fine
with me.  It may take me some time to go through the other drivers. I
don't know about retagging, that would be up to John I think. I would
say if you have a fix for coreaudio and I put in the call to drain for
pulseaudio, then might as well get those fixes in at least.

On Thu, Jan 21, 2010 at 6:29 PM, Victor Lazzarini
 wrote:
> Again, Jack is OK on OSX as I did the same test as Andres and did not
> see anything wrong. It seems OSX escaped most of the issues you
> reported (apart from the coreaudio module, but that is fixed now).
>
> Do you think I can move to build the binaries or should I wait? Are
> the sources going to be retagged?
>
> Victor
> On 21 Jan 2010, at 22:09, Steven Yi wrote:
>
>> I only have MME here for PA.  ASIO4ALL causes some weird issue on this
>> computer.  I think you're right though that it must have to do with
>> the underlying API.  Interesting to note the blocking API works with
>> MME.  I may take a look at what is going on in the portaudio blocking
>> for MME to fix up our MME driver.  I'll try to take a look at Jack
>> next as Andres reported that being an issue.
>>
>> On Thu, Jan 21, 2010 at 12:51 PM, Victor Lazzarini
>>  wrote:
>>> Really strange that it does not happen on OSX. It is probably the
>>> underlying API (mme, alsa)
>>> behind portaudio. Does it happen with ASIO or just with MME?
>>>
>>> Victor
>>>
>>> On 21 Jan 2010, at 17:40, Steven Yi wrote:
>>>
>>>> Hi All,
>>>>
>>>> I've fiddled with the portaudio driver callback closing this morning
>>>> and haven't managed to find a way to get it properly wait to close.
>>>> Strangely enough, I've added Pa_Sleep all around the closing code in
>>>> rtclose_ and it doesn't seem to remove the pop.  I've tried all
>>>> manner
>>>> of changes.  I'm going to give up on this for now and move to see
>>>> about removing pops in the other drivers.
>>>>
>>>> Thanks!
>>>> steven
>>>>
>>>> On Wed, Jan 20, 2010 at 11:46 PM, Steven Yi 
>>>> wrote:
>>>>> Adding the flush to pulseaudio fixed the issue for that driver
>>>>> (committed to CVS).  I am guessing at this point it is a problem
>>>>> that
>>>>> is in different drivers and we'll have to go and check each.  Seems
>>>>> like we should just go through all of the drivers and see if the
>>>>> audio
>>>>> API each driver is using has some means of draining or flushing out
>>>>> the buffer before closing, and if we're doing it correctly.
>>>>>
>>>>> On Wed, Jan 20, 2010 at 10:43 PM, Steven Yi 
>>>>> wrote:
>>>>>> Yeah, I'm thinking now though if that was a problem then it'd
>>>>>> affect
>>>>>> output to disk.  I should have thought earlier to remember that
>>>>>> since
>>>>>> one of the audio drivers is working here (pa_bl) then it would
>>>>>> mean
>>>>>> csound's code should be fine. I'm beginning to think that we just
>>>>>> need
>>>>>> to go through each audio driver and make sure it is flushing
>>>>>> properly
>>>>>> before closing.  I was experimenting with modifications to
>>>>>> rtwinmm.c's
>>>>>> rtclose and managed to hear more of the output though still
>>>>>> slightly
>>>>>> chopped off.  I also noticed the pulseaudio driver does not do a
>>>>>> flush
>>>>>> with pa_simple_drain() before close (going to try this out now on
>>>>>> Fedora on my netbook).
>>>>>>
>>>>>>
>>>>>> On Wed, Jan 20, 2010 at 9:48 PM,   wrote:
>>>>>>> The only change I can think of was the change from counting in
>>>>>>> seconds
>>>>>>> (floating point) to counting in samples  There could be a > or >=
>>>>>>> prpoblem
>>>>>>> there.  Mainly in musmon.c
>>>>>>>  That does not seem to be enough to explain i tthough
>>>>>>>
>>>>>>> ==John ff
>>>>>>>
>>>>>>>> Hi All,
>>>>>>>>
>>>>>>>> I'm looking at libsnd.c, particularly sfcloseout.  It has this
>>>>>>>> code:
>>>>>>>>
>>>>>>>>     if ((nb = (O->outbufsamps - ST(outbufrem)) * sizeof(MYFLT))
>>>>>>>>> 0) {
>>>>>>>>       /* flush outbuffer */
>>>>>>>>       csound->nrecs++;
>>>>>>>>       csound->audtran(csound, ST(outbuf), nb);
>>>>>>>>     }
>>>>>>>>
>>>>>>>> On this machine I had O->outbufsamps = 8192, and using this
>>>>>>>> before the
>>>>>>>> statement block above:
>>>>>>>>
>>>>>>>>     csound->Message(csound, ">>>> %d", ST(outbufrem));
>>>>>>>>
>>>>>>>> I got an outbufrem of 136854.    The code looks like this all
>>>>>>>> the way
>>>>>>>> back the first revision of libsnd.c in CVS but at that time
>>>>>>>> outbufrem
>>>>>>>> was a static variable and I assume was used elsewhere (in that
>>>>>>>> version
>>>>>>>> of the file, it is only used in this block of code, not set
>>>>>>>> anywhere).
>>>>>>>>  I am wondering if something here changed causing an issue?  I
>>>>>>>> also
>>>>>>>> saw that in the oldest code, it was:
>>>>>>>>
>>>>>>>>     if ((nb = (O.outbufsamps-outbufrem) *
>>>>>>>>          O.outsampsiz) > 0)/* flush outbuffer */
>>>>>>>>
>>>>>>>> using sample size, but it seemed to have changed to MYFLT at
>>>>>>>> some point.
>>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------------
>>>>>>> Throughout its 18-year history, RSA Conference consistently
>>>>>>> attracts the
>>>>>>> world's best and brightest in the field, creating opportunities
>>>>>>> for Conference
>>>>>>> attendees to learn about information security's most important
>>>>>>> issues through
>>>>>>> interactions with peers, luminaries and emerging and established
>>>>>>> companies.
>>>>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>>>>> _______________________________________________
>>>>>>> Csound-devel mailing list
>>>>>>> Csound-devel@lists.sourceforge.net
>>>>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>>>>>
>>>>>>
>>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>> Throughout its 18-year history, RSA Conference consistently attracts
>>>> the
>>>> world's best and brightest in the field, creating opportunities for
>>>> Conference
>>>> attendees to learn about information security's most important
>>>> issues through
>>>> interactions with peers, luminaries and emerging and established
>>>> companies.
>>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>>> _______________________________________________
>>>> Csound-devel mailing list
>>>> Csound-devel@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Throughout its 18-year history, RSA Conference consistently
>>> attracts the
>>> world's best and brightest in the field, creating opportunities for
>>> Conference
>>> attendees to learn about information security's most important
>>> issues through
>>> interactions with peers, luminaries and emerging and established
>>> companies.
>>> http://p.sf.net/sfu/rsaconf-dev2dev
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>>
>>
>> ------------------------------------------------------------------------------
>> Throughout its 18-year history, RSA Conference consistently attracts
>> the
>> world's best and brightest in the field, creating opportunities for
>> Conference
>> attendees to learn about information security's most important
>> issues through
>> interactions with peers, luminaries and emerging and established
>> companies.
>> http://p.sf.net/sfu/rsaconf-dev2dev
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
> ------------------------------------------------------------------------------
> Throughout its 18-year history, RSA Conference consistently attracts the
> world's best and brightest in the field, creating opportunities for Conference
> attendees to learn about information security's most important issues through
> interactions with peers, luminaries and emerging and established companies.
> http://p.sf.net/sfu/rsaconf-dev2dev
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>

------------------------------------------------------------------------------
Throughout its 18-year history, RSA Conference consistently attracts the
world's best and brightest in the field, creating opportunities for Conference
attendees to learn about information security's most important issues through
interactions with peers, luminaries and emerging and established companies.
http://p.sf.net/sfu/rsaconf-dev2dev
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net