Csound Csound-dev Csound-tekno Search About

[Cs-dev] FFT of a whole sound file (no realtime / no STFT)

Date2013-08-13 21:33
Frombaruchel@gmx.com
Subject[Cs-dev] FFT of a whole sound file (no realtime / no STFT)
Hi, it looks like it is very easy to perform STFT with csound, but
is it possible to perform a single FFT on a whole sound file, make some changes
on it, then reconstruct the whole sound? Regards, b.


------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 08:49
FromVictor Lazzarini
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
Theoretically you could set the fftsize to match the soundfile size, but I have not tried it and
I do not know if it works correctly. You should be able to do this with the pvanal utility 
and with the pvsanal opcode.

On 13 Aug 2013, at 21:33, baruchel@gmx.com wrote:

> 
> Hi, it looks like it is very easy to perform STFT with csound, but
> is it possible to perform a single FFT on a whole sound file, make some changes
> on it, then reconstruct the whole sound? Regards, b.
> 
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead. 
> Download for free and get started troubleshooting in minutes. 
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 10:05
FromVictor Lazzarini
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
It appears to work, I tried

instr 1
S1 = "pianoc2.wav"
ilen filelen S1
a1 diskin2 S1,1
fs1 pvsanal a1, ilen, ilen, ilen, 1
a2 pvsynth fs1
out a2,a2
endin

On 14 Aug 2013, at 08:49, Victor Lazzarini wrote:

> Theoretically you could set the fftsize to match the soundfile size, but I have not tried it and
> I do not know if it works correctly. You should be able to do this with the pvanal utility 
> and with the pvsanal opcode.
> 
> On 13 Aug 2013, at 21:33, baruchel@gmx.com wrote:
> 
>> 
>> Hi, it looks like it is very easy to perform STFT with csound, but
>> is it possible to perform a single FFT on a whole sound file, make some changes
>> on it, then reconstruct the whole sound? Regards, b.
>> 
>> 
>> ------------------------------------------------------------------------------
>> Get 100% visibility into Java/.NET code with AppDynamics Lite!
>> It's a free troubleshooting tool designed for production.
>> Get down to code-level detail for bottlenecks, with <2% overhead. 
>> Download for free and get started troubleshooting in minutes. 
>> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> Dr Victor Lazzarini
> Senior Lecturer
> Dept. of Music
> NUI Maynooth Ireland
> tel.: +353 1 708 3545
> Victor dot Lazzarini AT nuim dot ie
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead. 
> Download for free and get started troubleshooting in minutes. 
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 10:05
Frompeiman khosravi
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
AttachmentsNone  None  
Will this result in a complete loss of temporal evolution? 



On 14 August 2013 08:49, Victor Lazzarini <Victor.Lazzarini@nuim.ie> wrote:
Theoretically you could set the fftsize to match the soundfile size, but I have not tried it and
I do not know if it works correctly. You should be able to do this with the pvanal utility
and with the pvsanal opcode.

On 13 Aug 2013, at 21:33, baruchel@gmx.com wrote:

>
> Hi, it looks like it is very easy to perform STFT with csound, but
> is it possible to perform a single FFT on a whole sound file, make some changes
> on it, then reconstruct the whole sound? Regards, b.
>
>
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead.
> Download for free and get started troubleshooting in minutes.
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead.
Download for free and get started troubleshooting in minutes.
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-devel


Date2013-08-14 10:14
FromVictor Lazzarini
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
It depends on what you mean by this. Certainly you can't have time-varying effects. It is a time-invariant process.

Victor
On 14 Aug 2013, at 10:05, peiman khosravi wrote:

> Will this result in a complete loss of temporal evolution? 
> 
> 
> 
> 
> www.peimankhosravi.co.uk || RSS Feed || Concert News
> 
> 
> On 14 August 2013 08:49, Victor Lazzarini  wrote:
> Theoretically you could set the fftsize to match the soundfile size, but I have not tried it and
> I do not know if it works correctly. You should be able to do this with the pvanal utility
> and with the pvsanal opcode.
> 
> On 13 Aug 2013, at 21:33, baruchel@gmx.com wrote:
> 
> >
> > Hi, it looks like it is very easy to perform STFT with csound, but
> > is it possible to perform a single FFT on a whole sound file, make some changes
> > on it, then reconstruct the whole sound? Regards, b.
> >
> >
> > ------------------------------------------------------------------------------
> > Get 100% visibility into Java/.NET code with AppDynamics Lite!
> > It's a free troubleshooting tool designed for production.
> > Get down to code-level detail for bottlenecks, with <2% overhead.
> > Download for free and get started troubleshooting in minutes.
> > http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> > _______________________________________________
> > Csound-devel mailing list
> > Csound-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> Dr Victor Lazzarini
> Senior Lecturer
> Dept. of Music
> NUI Maynooth Ireland
> tel.: +353 1 708 3545
> Victor dot Lazzarini AT nuim dot ie
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead.
> Download for free and get started troubleshooting in minutes.
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead. 
> Download for free and get started troubleshooting in minutes. 
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk_______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 10:17
Frompeiman khosravi
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
AttachmentsNone  None  
I mean does one loos the temporal information in the input sound? I imagine that it would become a frozen spectrum.



On 14 August 2013 10:14, Victor Lazzarini <Victor.Lazzarini@nuim.ie> wrote:
It depends on what you mean by this. Certainly you can't have time-varying effects. It is a time-invariant process.

Victor
On 14 Aug 2013, at 10:05, peiman khosravi wrote:

> Will this result in a complete loss of temporal evolution?
>
>
>
>
> www.peimankhosravi.co.uk || RSS Feed || Concert News
>
>
> On 14 August 2013 08:49, Victor Lazzarini <Victor.Lazzarini@nuim.ie> wrote:
> Theoretically you could set the fftsize to match the soundfile size, but I have not tried it and
> I do not know if it works correctly. You should be able to do this with the pvanal utility
> and with the pvsanal opcode.
>
> On 13 Aug 2013, at 21:33, baruchel@gmx.com wrote:
>
> >
> > Hi, it looks like it is very easy to perform STFT with csound, but
> > is it possible to perform a single FFT on a whole sound file, make some changes
> > on it, then reconstruct the whole sound? Regards, b.
> >
> >
> > ------------------------------------------------------------------------------
> > Get 100% visibility into Java/.NET code with AppDynamics Lite!
> > It's a free troubleshooting tool designed for production.
> > Get down to code-level detail for bottlenecks, with <2% overhead.
> > Download for free and get started troubleshooting in minutes.
> > http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> > _______________________________________________
> > Csound-devel mailing list
> > Csound-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/csound-devel
>
> Dr Victor Lazzarini
> Senior Lecturer
> Dept. of Music
> NUI Maynooth Ireland
> tel.: +353 1 708 3545
> Victor dot Lazzarini AT nuim dot ie
>
>
>
>
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead.
> Download for free and get started troubleshooting in minutes.
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
>
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead.
> Download for free and get started troubleshooting in minutes.
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk_______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead.
Download for free and get started troubleshooting in minutes.
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-devel


Date2013-08-14 11:02
FromVictor Lazzarini
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
No, if there is no transformation, the output == input. The usual pvs transformations will probably not work,
as they depend on the inter-frame phase difference. So it may be a curiosity only.

Victor
On 14 Aug 2013, at 10:17, peiman khosravi wrote:

> I mean does one loos the temporal information in the input sound? I imagine that it would become a frozen spectrum.
> 
> 
> 
> 
> www.peimankhosravi.co.uk || RSS Feed || Concert News
> 
> 
> On 14 August 2013 10:14, Victor Lazzarini  wrote:
> It depends on what you mean by this. Certainly you can't have time-varying effects. It is a time-invariant process.
> 
> Victor
> On 14 Aug 2013, at 10:05, peiman khosravi wrote:
> 
> > Will this result in a complete loss of temporal evolution?
> >
> >
> >
> >
> > www.peimankhosravi.co.uk || RSS Feed || Concert News
> >
> >
> > On 14 August 2013 08:49, Victor Lazzarini  wrote:
> > Theoretically you could set the fftsize to match the soundfile size, but I have not tried it and
> > I do not know if it works correctly. You should be able to do this with the pvanal utility
> > and with the pvsanal opcode.
> >
> > On 13 Aug 2013, at 21:33, baruchel@gmx.com wrote:
> >
> > >
> > > Hi, it looks like it is very easy to perform STFT with csound, but
> > > is it possible to perform a single FFT on a whole sound file, make some changes
> > > on it, then reconstruct the whole sound? Regards, b.
> > >
> > >
> > > ------------------------------------------------------------------------------
> > > Get 100% visibility into Java/.NET code with AppDynamics Lite!
> > > It's a free troubleshooting tool designed for production.
> > > Get down to code-level detail for bottlenecks, with <2% overhead.
> > > Download for free and get started troubleshooting in minutes.
> > > http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> > > _______________________________________________
> > > Csound-devel mailing list
> > > Csound-devel@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/csound-devel
> >
> > Dr Victor Lazzarini
> > Senior Lecturer
> > Dept. of Music
> > NUI Maynooth Ireland
> > tel.: +353 1 708 3545
> > Victor dot Lazzarini AT nuim dot ie
> >
> >
> >
> >
> > ------------------------------------------------------------------------------
> > Get 100% visibility into Java/.NET code with AppDynamics Lite!
> > It's a free troubleshooting tool designed for production.
> > Get down to code-level detail for bottlenecks, with <2% overhead.
> > Download for free and get started troubleshooting in minutes.
> > http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> > _______________________________________________
> > Csound-devel mailing list
> > Csound-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/csound-devel
> >
> > ------------------------------------------------------------------------------
> > Get 100% visibility into Java/.NET code with AppDynamics Lite!
> > It's a free troubleshooting tool designed for production.
> > Get down to code-level detail for bottlenecks, with <2% overhead.
> > Download for free and get started troubleshooting in minutes.
> > http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk_______________________________________________
> > Csound-devel mailing list
> > Csound-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> Dr Victor Lazzarini
> Senior Lecturer
> Dept. of Music
> NUI Maynooth Ireland
> tel.: +353 1 708 3545
> Victor dot Lazzarini AT nuim dot ie
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead.
> Download for free and get started troubleshooting in minutes.
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead. 
> Download for free and get started troubleshooting in minutes. 
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk_______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 11:17
FromRichard Dobson
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
This is the idea behind the Mammut project:

http://www.notam02.no/web/2005/10/mammut-2/?lang=en

"
Doing a giant FFT of the entire sound, as opposed to splitting the sound 
up into short windows, is unusual. Such a method implies that 
time-related parameters are included in the spectral coefficients in a 
non-intuitive manner, and changes in the frequency domain may radically 
change developments in the time domain.
"

So yes, the pvoc approach is not really indicated here, it is 
specifically transformation on a single huge frame.

Note that CARL pvoc (as used by pvsanal) generates a minimum of two 
frames in an analysis file, where the first frame is always empty (zero 
values). The required approach is probably to write the analysis data to 
a table, and process that bin by bin, somehow.


Richard Dobson



On 14/08/2013 11:02, Victor Lazzarini wrote:
> No, if there is no transformation, the output == input. The usual pvs
> transformations will probably not work, as they depend on the
> inter-frame phase difference. So it may be a curiosity only.
>

------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 11:51
FromVictor Lazzarini
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
yes, the PVOC approach is not very useful here, but the current pvsanal implementation
appears to allow a single frame to be analysed and resynthesised with no transformations,
in a transparent way (or at least it sounds that way). I have not looked at the code looking to confirm
this, but experimentally it does appear that way.

On 14 Aug 2013, at 11:17, Richard Dobson wrote:

> Note that CARL pvoc (as used by pvsanal) generates a minimum of two 
> frames in an analysis file, where the first frame is always empty (zero 
> values). The required approach is probably to write the analysis data to 
> a table, and process that bin by bin, somehow.

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 11:57
FromVictor Lazzarini
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
Sorry, this is actually wrong. The fftsize should be ilen*sr, but I think there is a size limit for it.
On 14 Aug 2013, at 10:05, Victor Lazzarini wrote:

> It appears to work, I tried
> 
> instr 1
> S1 = "pianoc2.wav"
> ilen filelen S1
> a1 diskin2 S1,1
> fs1 pvsanal a1, ilen, ilen, ilen, 1
> a2 pvsynth fs1
> out a2,a2
> endin
> 
> On 14 Aug 2013, at 08:49, Victor Lazzarini wrote:
> 
>> Theoretically you could set the fftsize to match the soundfile size, but I have not tried it and
>> I do not know if it works correctly. You should be able to do this with the pvanal utility 
>> and with the pvsanal opcode.
>> 
>> On 13 Aug 2013, at 21:33, baruchel@gmx.com wrote:
>> 
>>> 
>>> Hi, it looks like it is very easy to perform STFT with csound, but
>>> is it possible to perform a single FFT on a whole sound file, make some changes
>>> on it, then reconstruct the whole sound? Regards, b.
>>> 
>>> 
>>> ------------------------------------------------------------------------------
>>> Get 100% visibility into Java/.NET code with AppDynamics Lite!
>>> It's a free troubleshooting tool designed for production.
>>> Get down to code-level detail for bottlenecks, with <2% overhead. 
>>> Download for free and get started troubleshooting in minutes. 
>>> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> 
>> Dr Victor Lazzarini
>> Senior Lecturer
>> Dept. of Music
>> NUI Maynooth Ireland
>> tel.: +353 1 708 3545
>> Victor dot Lazzarini AT nuim dot ie
>> 
>> 
>> 
>> 
>> ------------------------------------------------------------------------------
>> Get 100% visibility into Java/.NET code with AppDynamics Lite!
>> It's a free troubleshooting tool designed for production.
>> Get down to code-level detail for bottlenecks, with <2% overhead. 
>> Download for free and get started troubleshooting in minutes. 
>> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> Dr Victor Lazzarini
> Senior Lecturer
> Dept. of Music
> NUI Maynooth Ireland
> tel.: +353 1 708 3545
> Victor dot Lazzarini AT nuim dot ie
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead. 
> Download for free and get started troubleshooting in minutes. 
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 11:57
FromVictor Lazzarini
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
My test was wrong, so forget what I said ;)
On 14 Aug 2013, at 11:51, Victor Lazzarini wrote:

> yes, the PVOC approach is not very useful here, but the current pvsanal implementation
> appears to allow a single frame to be analysed and resynthesised with no transformations,
> in a transparent way (or at least it sounds that way). I have not looked at the code looking to confirm
> this, but experimentally it does appear that way.
> 
> On 14 Aug 2013, at 11:17, Richard Dobson wrote:
> 
>> Note that CARL pvoc (as used by pvsanal) generates a minimum of two 
>> frames in an analysis file, where the first frame is always empty (zero 
>> values). The required approach is probably to write the analysis data to 
>> a table, and process that bin by bin, somehow.
> 
> Dr Victor Lazzarini
> Senior Lecturer
> Dept. of Music
> NUI Maynooth Ireland
> tel.: +353 1 708 3545
> Victor dot Lazzarini AT nuim dot ie
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead. 
> Download for free and get started troubleshooting in minutes. 
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 12:00
Frompeiman khosravi
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
AttachmentsNone  None  
Thanks Victor. I wonder what would happen if one started to mess with the data. Time to experiment. 

P



On 14 August 2013 11:51, Victor Lazzarini <Victor.Lazzarini@nuim.ie> wrote:
yes, the PVOC approach is not very useful here, but the current pvsanal implementation
appears to allow a single frame to be analysed and resynthesised with no transformations,
in a transparent way (or at least it sounds that way). I have not looked at the code looking to confirm
this, but experimentally it does appear that way.

On 14 Aug 2013, at 11:17, Richard Dobson wrote:

> Note that CARL pvoc (as used by pvsanal) generates a minimum of two
> frames in an analysis file, where the first frame is always empty (zero
> values). The required approach is probably to write the analysis data to
> a table, and process that bin by bin, somehow.

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead.
Download for free and get started troubleshooting in minutes.
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-devel


Date2013-08-14 12:02
Frompeiman khosravi
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
AttachmentsNone  None  
Oh well...

This thread makes me nervous: I need to keep developing FFTools but csound6 is not yet supported by csound~.

 



On 14 August 2013 12:00, peiman khosravi <peimankhosravi@gmail.com> wrote:
Thanks Victor. I wonder what would happen if one started to mess with the data. Time to experiment. 

P
On 14 August 2013 11:51, Victor Lazzarini <Victor.Lazzarini@nuim.ie> wrote:
yes, the PVOC approach is not very useful here, but the current pvsanal implementation
appears to allow a single frame to be analysed and resynthesised with no transformations,
in a transparent way (or at least it sounds that way). I have not looked at the code looking to confirm
this, but experimentally it does appear that way.

On 14 Aug 2013, at 11:17, Richard Dobson wrote:

> Note that CARL pvoc (as used by pvsanal) generates a minimum of two
> frames in an analysis file, where the first frame is always empty (zero
> values). The required approach is probably to write the analysis data to
> a table, and process that bin by bin, somehow.

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead.
Download for free and get started troubleshooting in minutes.
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-devel



Date2013-08-14 12:06
FromVictor Lazzarini
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
I looked at the code and indeed it requires the overlap to be at least N/2

On 14 Aug 2013, at 11:57, Victor Lazzarini wrote:

> My test was wrong, so forget what I said ;)
> On 14 Aug 2013, at 11:51, Victor Lazzarini wrote:
> 
>> yes, the PVOC approach is not very useful here, but the current pvsanal implementation
>> appears to allow a single frame to be analysed and resynthesised with no transformations,
>> in a transparent way (or at least it sounds that way). I have not looked at the code looking to confirm
>> this, but experimentally it does appear that way.
>> 
>> On 14 Aug 2013, at 11:17, Richard Dobson wrote:
>> 
>>> Note that CARL pvoc (as used by pvsanal) generates a minimum of two 
>>> frames in an analysis file, where the first frame is always empty (zero 
>>> values). The required approach is probably to write the analysis data to 
>>> a table, and process that bin by bin, somehow.
>> 
>> Dr Victor Lazzarini
>> Senior Lecturer
>> Dept. of Music
>> NUI Maynooth Ireland
>> tel.: +353 1 708 3545
>> Victor dot Lazzarini AT nuim dot ie
>> 
>> 
>> 
>> 
>> ------------------------------------------------------------------------------
>> Get 100% visibility into Java/.NET code with AppDynamics Lite!
>> It's a free troubleshooting tool designed for production.
>> Get down to code-level detail for bottlenecks, with <2% overhead. 
>> Download for free and get started troubleshooting in minutes. 
>> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> Dr Victor Lazzarini
> Senior Lecturer
> Dept. of Music
> NUI Maynooth Ireland
> tel.: +353 1 708 3545
> Victor dot Lazzarini AT nuim dot ie
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead. 
> Download for free and get started troubleshooting in minutes. 
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 12:26
FromRichard Dobson
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
Maybe something can be done via cvanal? That by definition produces a 
single analysis frame. AFAIK at present we only have the convolve opcode 
to use with it, but if the data can be accessed directly as a table, it 
could be played around with ad lib.

Richard Dobson

On 14/08/2013 12:06, Victor Lazzarini wrote:
> I looked at the code and indeed it requires the overlap to be at least N/2
>


------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Date2013-08-14 12:44
FromVictor Lazzarini
SubjectRe: [Cs-dev] FFT of a whole sound file (no realtime / no STFT)
After correcting my example, this works:

instr 1
S1 = "pianoc2.wav"
ilen filelen S1
a1 diskin2 S1,1
ilen = ilen*sr
fs1 pvsanal a1, ilen, ilen/2, ilen, 1
a2 pvsynth fs1
out a2,a2
endin


so we need a 1/2 frame overlap, but we can do the whole file. There is a big delay
between input and output, though. Because it's a monster of an FFT, there is a big
surge in CPU demand at a given point, causing a click, but otherwise it runs quite
fast.

Victor

On 14 Aug 2013, at 12:06, Victor Lazzarini wrote:

> I looked at the code and indeed it requires the overlap to be at least N/2
> 
> On 14 Aug 2013, at 11:57, Victor Lazzarini wrote:
> 
>> My test was wrong, so forget what I said ;)
>> On 14 Aug 2013, at 11:51, Victor Lazzarini wrote:
>> 
>>> yes, the PVOC approach is not very useful here, but the current pvsanal implementation
>>> appears to allow a single frame to be analysed and resynthesised with no transformations,
>>> in a transparent way (or at least it sounds that way). I have not looked at the code looking to confirm
>>> this, but experimentally it does appear that way.
>>> 
>>> On 14 Aug 2013, at 11:17, Richard Dobson wrote:
>>> 
>>>> Note that CARL pvoc (as used by pvsanal) generates a minimum of two 
>>>> frames in an analysis file, where the first frame is always empty (zero 
>>>> values). The required approach is probably to write the analysis data to 
>>>> a table, and process that bin by bin, somehow.
>>> 
>>> Dr Victor Lazzarini
>>> Senior Lecturer
>>> Dept. of Music
>>> NUI Maynooth Ireland
>>> tel.: +353 1 708 3545
>>> Victor dot Lazzarini AT nuim dot ie
>>> 
>>> 
>>> 
>>> 
>>> ------------------------------------------------------------------------------
>>> Get 100% visibility into Java/.NET code with AppDynamics Lite!
>>> It's a free troubleshooting tool designed for production.
>>> Get down to code-level detail for bottlenecks, with <2% overhead. 
>>> Download for free and get started troubleshooting in minutes. 
>>> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
>>> _______________________________________________
>>> Csound-devel mailing list
>>> Csound-devel@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> 
>> Dr Victor Lazzarini
>> Senior Lecturer
>> Dept. of Music
>> NUI Maynooth Ireland
>> tel.: +353 1 708 3545
>> Victor dot Lazzarini AT nuim dot ie
>> 
>> 
>> 
>> 
>> ------------------------------------------------------------------------------
>> Get 100% visibility into Java/.NET code with AppDynamics Lite!
>> It's a free troubleshooting tool designed for production.
>> Get down to code-level detail for bottlenecks, with <2% overhead. 
>> Download for free and get started troubleshooting in minutes. 
>> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> Dr Victor Lazzarini
> Senior Lecturer
> Dept. of Music
> NUI Maynooth Ireland
> tel.: +353 1 708 3545
> Victor dot Lazzarini AT nuim dot ie
> 
> 
> 
> 
> ------------------------------------------------------------------------------
> Get 100% visibility into Java/.NET code with AppDynamics Lite!
> It's a free troubleshooting tool designed for production.
> Get down to code-level detail for bottlenecks, with <2% overhead. 
> Download for free and get started troubleshooting in minutes. 
> http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel

Dr Victor Lazzarini
Senior Lecturer
Dept. of Music
NUI Maynooth Ireland
tel.: +353 1 708 3545
Victor dot Lazzarini AT nuim dot ie




------------------------------------------------------------------------------
Get 100% visibility into Java/.NET code with AppDynamics Lite!
It's a free troubleshooting tool designed for production.
Get down to code-level detail for bottlenecks, with <2% overhead. 
Download for free and get started troubleshooting in minutes. 
http://pubads.g.doubleclick.net/gampad/clk?id=48897031&iu=/4140/ostg.clktrk
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net