Has anyone done oversampling (in RT) in Csound?,
say, to do alias-free distortion etc....
Would it be possible with standard opcodes?
...writing to table or array, interpolating, processing (distortion),
decimation, output.

You can do it in an UDO (with local ksmps = 1)
For instance, I do very simple 2x oversampling in the UDO version of the moogladder filter,
by running the filter twice on each sample. But other approaches could be taken, I suppose, since
Csound allows to process audio sample by sample, and you can write straight expressions 
to do it. I would expect that you might need some sort of buffering to upsample each sample input, 
process that data in maybe a loop, then downsample into each sample for output. 

best regards
========================
Dr Victor Lazzarini
Dean of Arts, Celtic Studies and Philosophy,
Maynooth University,
Maynooth, Co Kildare, Ireland
Tel: 00 353 7086936
Fax: 00 353 1 7086952 

> On 24 May 2015, at 21:55, Oeyvind Brandtsegg  wrote:
> 
> Has anyone done oversampling (in RT) in Csound?,
> say, to do alias-free distortion etc....
> Would it be possible with standard opcodes?
> ...writing to table or array, interpolating, processing (distortion),
> decimation, output.
> 
> -- 
> 
> Oeyvind Brandtsegg
> Professor of Music Technology
> NTNU
> 7491 Trondheim
> Norway
> Cell: +47 92 203 205
> 
> http://flyndresang.no/
> http://www.partikkelaudio.com/
> http://soundcloud.com/brandtsegg
> http://soundcloud.com/t-emp
> 
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud 
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> Csound-users mailing list
> Csound-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-users
> Send bugs reports to
>        https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here


------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Csound-users mailing list
Csound-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-users
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

I've been wondering for a while if the UDO mechanism couldn't support
this itself. I imagine some kind of statement at the beginning and end
of the UDO block could signal the sampling ratio and then it could
instantiate appropriate interpolation and decimation filters for the
sample rate and take care of any delay as a result of filtering. Then
all of the code in the UDO could remain unchanged. I guess it depends
on how the UDO works internally and on whether the block processing is
amenable to that.

Not being very familiar with csounds internals, I wouldn't know where
to look but if anyone can point me in the right direction, I'd be
interested in looking into this.

Regards,
 Joel

On 25 May 2015 at 10:23, Victor Lazzarini  wrote:
> You can do it in an UDO (with local ksmps = 1)
> For instance, I do very simple 2x oversampling in the UDO version of the moogladder filter,
> by running the filter twice on each sample. But other approaches could be taken, I suppose, since
> Csound allows to process audio sample by sample, and you can write straight expressions
> to do it. I would expect that you might need some sort of buffering to upsample each sample input,
> process that data in maybe a loop, then downsample into each sample for output.
>
> best regards
> ========================
> Dr Victor Lazzarini
> Dean of Arts, Celtic Studies and Philosophy,
> Maynooth University,
> Maynooth, Co Kildare, Ireland
> Tel: 00 353 7086936
> Fax: 00 353 1 7086952
>
>> On 24 May 2015, at 21:55, Oeyvind Brandtsegg  wrote:
>>
>> Has anyone done oversampling (in RT) in Csound?,
>> say, to do alias-free distortion etc....
>> Would it be possible with standard opcodes?
>> ...writing to table or array, interpolating, processing (distortion),
>> decimation, output.
>>
>> --
>>
>> Oeyvind Brandtsegg
>> Professor of Music Technology
>> NTNU
>> 7491 Trondheim
>> Norway
>> Cell: +47 92 203 205
>>
>> http://flyndresang.no/
>> http://www.partikkelaudio.com/
>> http://soundcloud.com/brandtsegg
>> http://soundcloud.com/t-emp
>>
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM Insight.
>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>> _______________________________________________
>> Csound-users mailing list
>> Csound-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-users
>> Send bugs reports to
>>        https://github.com/csound/csound/issues
>> Discussions of bugs and features can be posted here
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> Csound-users mailing list
> Csound-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-users
> Send bugs reports to
>         https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Csound-users mailing list
Csound-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-users
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

That might be possible, but for now, it is still possible to do it "by hand".

Victor Lazzarini
Dean of Arts, Celtic Studies, and Philosophy
Maynooth University
Ireland

> On 25 May 2015, at 11:09, Joel Ross  wrote:
> 
> I've been wondering for a while if the UDO mechanism couldn't support
> this itself. I imagine some kind of statement at the beginning and end
> of the UDO block could signal the sampling ratio and then it could
> instantiate appropriate interpolation and decimation filters for the
> sample rate and take care of any delay as a result of filtering. Then
> all of the code in the UDO could remain unchanged. I guess it depends
> on how the UDO works internally and on whether the block processing is
> amenable to that.
> 
> Not being very familiar with csounds internals, I wouldn't know where
> to look but if anyone can point me in the right direction, I'd be
> interested in looking into this.
> 
> Regards,
> Joel
> 
>> On 25 May 2015 at 10:23, Victor Lazzarini  wrote:
>> You can do it in an UDO (with local ksmps = 1)
>> For instance, I do very simple 2x oversampling in the UDO version of the moogladder filter,
>> by running the filter twice on each sample. But other approaches could be taken, I suppose, since
>> Csound allows to process audio sample by sample, and you can write straight expressions
>> to do it. I would expect that you might need some sort of buffering to upsample each sample input,
>> process that data in maybe a loop, then downsample into each sample for output.
>> 
>> best regards
>> ========================
>> Dr Victor Lazzarini
>> Dean of Arts, Celtic Studies and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>> 
>>> On 24 May 2015, at 21:55, Oeyvind Brandtsegg  wrote:
>>> 
>>> Has anyone done oversampling (in RT) in Csound?,
>>> say, to do alias-free distortion etc....
>>> Would it be possible with standard opcodes?
>>> ...writing to table or array, interpolating, processing (distortion),
>>> decimation, output.
>>> 
>>> --
>>> 
>>> Oeyvind Brandtsegg
>>> Professor of Music Technology
>>> NTNU
>>> 7491 Trondheim
>>> Norway
>>> Cell: +47 92 203 205
>>> 
>>> http://flyndresang.no/
>>> http://www.partikkelaudio.com/
>>> http://soundcloud.com/brandtsegg
>>> http://soundcloud.com/t-emp
>>> 
>>> ------------------------------------------------------------------------------
>>> One dashboard for servers and applications across Physical-Virtual-Cloud
>>> Widest out-of-the-box monitoring support with 50+ applications
>>> Performance metrics, stats and reports that give you Actionable Insights
>>> Deep dive visibility with transaction tracing using APM Insight.
>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>> _______________________________________________
>>> Csound-users mailing list
>>> Csound-users@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-users
>>> Send bugs reports to
>>>       https://github.com/csound/csound/issues
>>> Discussions of bugs and features can be posted here
>> 
>> 
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM Insight.
>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>> _______________________________________________
>> Csound-users mailing list
>> Csound-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-users
>> Send bugs reports to
>>        https://github.com/csound/csound/issues
>> Discussions of bugs and features can be posted here
> 
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud 
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> Csound-users mailing list
> Csound-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-users
> Send bugs reports to
>        https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Csound-users mailing list
Csound-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-users
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

Hi Joel,

I've thought about a generic way to do oversampling as well but wasn't
quite sure about how to make the system generic enough for
choosing/implementing how to upsample/downsample.

As Victor mentioned, it's entirely possible to implement oversampling
in ORC code.  The nice thing would be if you could do something like:

setksmps(1)
setoversampling(4)

For oversampling, because opcodes often take into account sr during
init phase, one would have to update the system so that opcodes read
sr from a "local sr", in the same way that ksmps is read from "local
ksmps".  I'd have to check, but it might be the case that it's already
being read from local sr.  Also, I think a macro is used for most
opcodes reading sr such that one change to the macro would update all
the opcodes.

One a local sr is available, the other part would be the upsampling of
inputs and downsampling of outputs.  That would have to wrap the
existing code to run instruments/UDO's.  The code is in
Engine/csound.c for instruments; check out kperf_nodebug() and
kperf_debug().  For UDO's and subinstruments, check out
Engine/insert.c at the useropcd1(), useropcd2(), and subinstr()
functions. You'll see code related to ids and pds (or IDS or PDS),
which are the init-time and perf-time chains of opcodes that get
walked in a while-loop.

That also brings up one other thing, which would be the need for local
spin and spout.  These are the places where audio in and audio out are
read from/written to.  We would need those to be local, so that the
local buffer could be x times larger in size for the oversampling.

I think that covers the things necessary.  If you look into it and
have further questions, please feel free to ask.  As some of this gets
into internal details, maybe further questions would be better on the
developer mailing list.

Cheers!
steven



On Mon, May 25, 2015 at 6:09 AM, Joel Ross  wrote:
> I've been wondering for a while if the UDO mechanism couldn't support
> this itself. I imagine some kind of statement at the beginning and end
> of the UDO block could signal the sampling ratio and then it could
> instantiate appropriate interpolation and decimation filters for the
> sample rate and take care of any delay as a result of filtering. Then
> all of the code in the UDO could remain unchanged. I guess it depends
> on how the UDO works internally and on whether the block processing is
> amenable to that.
>
> Not being very familiar with csounds internals, I wouldn't know where
> to look but if anyone can point me in the right direction, I'd be
> interested in looking into this.
>
> Regards,
>  Joel
>
> On 25 May 2015 at 10:23, Victor Lazzarini  wrote:
>> You can do it in an UDO (with local ksmps = 1)
>> For instance, I do very simple 2x oversampling in the UDO version of the moogladder filter,
>> by running the filter twice on each sample. But other approaches could be taken, I suppose, since
>> Csound allows to process audio sample by sample, and you can write straight expressions
>> to do it. I would expect that you might need some sort of buffering to upsample each sample input,
>> process that data in maybe a loop, then downsample into each sample for output.
>>
>> best regards
>> ========================
>> Dr Victor Lazzarini
>> Dean of Arts, Celtic Studies and Philosophy,
>> Maynooth University,
>> Maynooth, Co Kildare, Ireland
>> Tel: 00 353 7086936
>> Fax: 00 353 1 7086952
>>
>>> On 24 May 2015, at 21:55, Oeyvind Brandtsegg  wrote:
>>>
>>> Has anyone done oversampling (in RT) in Csound?,
>>> say, to do alias-free distortion etc....
>>> Would it be possible with standard opcodes?
>>> ...writing to table or array, interpolating, processing (distortion),
>>> decimation, output.
>>>
>>> --
>>>
>>> Oeyvind Brandtsegg
>>> Professor of Music Technology
>>> NTNU
>>> 7491 Trondheim
>>> Norway
>>> Cell: +47 92 203 205
>>>
>>> http://flyndresang.no/
>>> http://www.partikkelaudio.com/
>>> http://soundcloud.com/brandtsegg
>>> http://soundcloud.com/t-emp
>>>
>>> ------------------------------------------------------------------------------
>>> One dashboard for servers and applications across Physical-Virtual-Cloud
>>> Widest out-of-the-box monitoring support with 50+ applications
>>> Performance metrics, stats and reports that give you Actionable Insights
>>> Deep dive visibility with transaction tracing using APM Insight.
>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>> _______________________________________________
>>> Csound-users mailing list
>>> Csound-users@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-users
>>> Send bugs reports to
>>>        https://github.com/csound/csound/issues
>>> Discussions of bugs and features can be posted here
>>
>>
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM Insight.
>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>> _______________________________________________
>> Csound-users mailing list
>> Csound-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-users
>> Send bugs reports to
>>         https://github.com/csound/csound/issues
>> Discussions of bugs and features can be posted here
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> Csound-users mailing list
> Csound-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-users
> Send bugs reports to
>         https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here

------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud 
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Csound-users mailing list
Csound-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/csound-users
Send bugs reports to
        https://github.com/csound/csound/issues
Discussions of bugs and features can be posted here

Excellent, thanks for all the replies.

2015-05-25 16:46 GMT+02:00 Steven Yi :
> Hi Joel,
>
> I've thought about a generic way to do oversampling as well but wasn't
> quite sure about how to make the system generic enough for
> choosing/implementing how to upsample/downsample.
>
> As Victor mentioned, it's entirely possible to implement oversampling
> in ORC code.  The nice thing would be if you could do something like:
>
> setksmps(1)
> setoversampling(4)
>
> For oversampling, because opcodes often take into account sr during
> init phase, one would have to update the system so that opcodes read
> sr from a "local sr", in the same way that ksmps is read from "local
> ksmps".  I'd have to check, but it might be the case that it's already
> being read from local sr.  Also, I think a macro is used for most
> opcodes reading sr such that one change to the macro would update all
> the opcodes.
>
> One a local sr is available, the other part would be the upsampling of
> inputs and downsampling of outputs.  That would have to wrap the
> existing code to run instruments/UDO's.  The code is in
> Engine/csound.c for instruments; check out kperf_nodebug() and
> kperf_debug().  For UDO's and subinstruments, check out
> Engine/insert.c at the useropcd1(), useropcd2(), and subinstr()
> functions. You'll see code related to ids and pds (or IDS or PDS),
> which are the init-time and perf-time chains of opcodes that get
> walked in a while-loop.
>
> That also brings up one other thing, which would be the need for local
> spin and spout.  These are the places where audio in and audio out are
> read from/written to.  We would need those to be local, so that the
> local buffer could be x times larger in size for the oversampling.
>
> I think that covers the things necessary.  If you look into it and
> have further questions, please feel free to ask.  As some of this gets
> into internal details, maybe further questions would be better on the
> developer mailing list.
>
> Cheers!
> steven
>
>
>
> On Mon, May 25, 2015 at 6:09 AM, Joel Ross  wrote:
>> I've been wondering for a while if the UDO mechanism couldn't support
>> this itself. I imagine some kind of statement at the beginning and end
>> of the UDO block could signal the sampling ratio and then it could
>> instantiate appropriate interpolation and decimation filters for the
>> sample rate and take care of any delay as a result of filtering. Then
>> all of the code in the UDO could remain unchanged. I guess it depends
>> on how the UDO works internally and on whether the block processing is
>> amenable to that.
>>
>> Not being very familiar with csounds internals, I wouldn't know where
>> to look but if anyone can point me in the right direction, I'd be
>> interested in looking into this.
>>
>> Regards,
>>  Joel
>>
>> On 25 May 2015 at 10:23, Victor Lazzarini  wrote:
>>> You can do it in an UDO (with local ksmps = 1)
>>> For instance, I do very simple 2x oversampling in the UDO version of the moogladder filter,
>>> by running the filter twice on each sample. But other approaches could be taken, I suppose, since
>>> Csound allows to process audio sample by sample, and you can write straight expressions
>>> to do it. I would expect that you might need some sort of buffering to upsample each sample input,
>>> process that data in maybe a loop, then downsample into each sample for output.
>>>
>>> best regards
>>> ========================
>>> Dr Victor Lazzarini
>>> Dean of Arts, Celtic Studies and Philosophy,
>>> Maynooth University,
>>> Maynooth, Co Kildare, Ireland
>>> Tel: 00 353 7086936
>>> Fax: 00 353 1 7086952
>>>
>>>> On 24 May 2015, at 21:55, Oeyvind Brandtsegg  wrote:
>>>>
>>>> Has anyone done oversampling (in RT) in Csound?,
>>>> say, to do alias-free distortion etc....
>>>> Would it be possible with standard opcodes?
>>>> ...writing to table or array, interpolating, processing (distortion),
>>>> decimation, output.
>>>>
>>>> --
>>>>
>>>> Oeyvind Brandtsegg
>>>> Professor of Music Technology
>>>> NTNU
>>>> 7491 Trondheim
>>>> Norway
>>>> Cell: +47 92 203 205
>>>>
>>>> http://flyndresang.no/
>>>> http://www.partikkelaudio.com/
>>>> http://soundcloud.com/brandtsegg
>>>> http://soundcloud.com/t-emp
>>>>
>>>> ------------------------------------------------------------------------------
>>>> One dashboard for servers and applications across Physical-Virtual-Cloud
>>>> Widest out-of-the-box monitoring support with 50+ applications
>>>> Performance metrics, stats and reports that give you Actionable Insights
>>>> Deep dive visibility with transaction tracing using APM Insight.
>>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>>> _______________________________________________
>>>> Csound-users mailing list
>>>> Csound-users@lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/csound-users
>>>> Send bugs reports to
>>>>        https://github.com/csound/csound/issues
>>>> Discussions of bugs and features can be posted here
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> One dashboard for servers and applications across Physical-Virtual-Cloud
>>> Widest out-of-the-box monitoring support with 50+ applications
>>> Performance metrics, stats and reports that give you Actionable Insights
>>> Deep dive visibility with transaction tracing using APM Insight.
>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>> _______________________________________________
>>> Csound-users mailing list
>>> Csound-users@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/csound-users
>>> Send bugs reports to
>>>         https://github.com/csound/csound/issues
>>> Discussions of bugs and features can be posted here
>>
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM Insight.
>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>> _______________________________________________
>> Csound-users mailing list
>> Csound-users@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-users
>> Send bugs reports to
>>         https://github.com/csound/csound/issues
>> Discussions of bugs and features can be posted here
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> Csound-users mailing list
> Csound-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-users
> Send bugs reports to
>         https://github.com/csound/csound/issues
> Discussions of bugs and features can be posted here