I think I have now fixed the specwarp issues in the pvoc opcodes.
The problem was probably also affecting pvcross and vpvoc, since
the same code was being used. There were two separate bugs,
responsible for the two funny behaviours detected by Oeyvind
(thanks for spotting and reporting them):

(1) a global spectral envelope was being used for all instances of
pvoc, pvcross and vpvoc, regardless of their sizes etc. This was
causing the interference between instruments
(2) the spectral warping function was accessing memory being
the allocated portion. This was causing the memory exception on
exit.

I have committed these fixes.

(By the way, the spectral warping works quite well, it's actually
very good. I'll look to see if I can implement a similar feature on
pvs opcodes. There might be some nice use I can put that bit of
code to).

Regards

Victor
----- Original Message ----- 
From: "Oeyvind Brandtsegg" 
To: "Developer discussions" 
Sent: Sunday, May 18, 2008 10:35 PM
Subject: Re: [Cs-dev] Responses from bug trawl


> Yes, it was for ispecw=1 only.
> However, to complicate matters, I have intermittently been able to run
> it without the bug appearing. If it runs without problems, it will run
> ok for repeated testing. If it does not run (happens more often), then
> trying it again will still not run it.
> ...sorry to be troublesome.
>
> Oeyvind
>
> 2008/5/18, victor :
>> I was able to run both instruments together (with ispecw=0). Is the
>> second bug (the one where you can't run more than one instrument)
>> with ispecw=1 only? (The only change I made was to use my own
>> pvx, the same for both instruments).
>>
>> Victor
>> ----- Original Message -----
>> From: "Oeyvind Brandtsegg" 
>> To: "Developer discussions" 
>> Sent: Sunday, May 18, 2008 9:44 PM
>> Subject: Re: [Cs-dev] Responses from bug trawl
>>
>>
>> >I do agree that the recommendation should be to use the pvs opcodes
>> >instead.
>> >
>> > I did supply a csd with the previous post (pvoc_test.csd),
>> > should I resend it ?
>> >
>> > best
>> > Oeyvind
>> >
>> >
>> > 2008/5/18, victor :
>> >> That needs to be looked after. Could you supply a CSD?
>> >>
>> >> I am wondering now whether we should mark the pvoc opcodes
>> >> as deprecated. Their functionality is replicated with much better
>> >> quality and flexibility by the pvs* opcodes.
>> >>
>> >> Victor
>> >> ----- Original Message -----
>> >> From: "Oeyvind Brandtsegg" 
>> >> To: "Developer discussions" 
>> >> Sent: Sunday, May 18, 2008 7:31 AM
>> >> Subject: Re: [Cs-dev] Responses from bug trawl
>> >>
>> >>
>> >> > And, sorry again, one more bug.
>> >> >
>> >> > The pvoc opcode seems to not work correclty if the ispecwp parameter
>> >> > is set to 1.
>> >> > For example
>> >> > a1 pvoc ktimpnt, 1.5, Spfile, 1
>> >> > will make sound, but also will print a
>> >> > "Csound tidy up: Segmentation violation"
>> >> > at the end of performance.
>> >> >
>> >> > That's a straightforward bug, but there's also one pvoc bug that 
>> >> > seem
>> >> > very strange indeed:
>> >> >
>> >> > Using two instrument containing a pvoc opcode in the same orchestra
>> >> > does not seem to work. None of the two instruments seem to make any
>> >> > sound, even if no warnings are printed. This bug shows up
>> >> > consistently, even if only one of the instruments is activated (see
>> >> > attached example).
>> >> > Intermittently, I have also been able to provoke the bug by just
>> >> > having another instrument defined in addition to the pvoc one. For
>> >> > example, in the attached file, replace all of instrument 2 with :
>> >> > instr 2
>> >> > itest = 0
>> >> > endin
>> >> >
>> >> > The itest line seems to be needed to provoke the bug (any text will 
>> >> > do,
>> >> > but just
>> >> > instr 2
>> >> > endin
>> >> > will not provoke the bug.
>> >> >
>> >> > This was tested (and the example made) on windows,
>> >> > but a student reported similar behaviour on OSX earlier this fall.
>> >> >
>> >> > Oeyvind
>> >> >
>> >> > 2008/5/18, Oeyvind Brandtsegg :
>> >> >> Sorry to chime in so late.
>> >> >> There is a bug related to using strings as p-fields.
>> >> >> If using more then one string p-field, the last one will also be 
>> >> >> used
>> >> >> as the string for any other (earlier) string p-fields for the same
>> >> >> note event.
>> >> >>
>> >> >> Example attached.
>> >> >>
>> >> >> Oeyvind
>> >> >>
>> >> >> 2008/5/17, jpff :
>> >> >> > This is a collection of the bugs reported and some responses
>> >> >> >
>> >> >> > * 'adsyn' and friends are non-functional.
>> >> >> >
>> >> >> >  I have started to look at LPANAL.I think we need better guidance
>> >> >> >
>> >> >> >
>> >> >> > * Python extension modules are built, but not installed (except
>> >> >> >  by hand)
>> >> >> >
>> >> >> >  Anyone?
>> >> >> >
>> >> >> > * STK opcodes compile from source, but Loris fail (they seem to
>> >> >> >  want a file that doesn't exist).
>> >> >> >
>> >> >> >  I did not understand this one I build Loris from source
>> >> >> >
>> >> >> > * Csound GUI on os x
>> >> >> >
>> >> >> >  Anyone?  Who is responsible for Csound GUI?  Or is this generic? 
>> >> >> > I
>> >> >> >  rather gave up on Winsound when others moved into that area
>> >> >> >
>> >> >> > * ATS opcodes don't work so well
>> >> >> >
>> >> >> >  Again we need evidence and examples
>> >> >> >
>> >> >> > * Loris on OS X -- and documentation
>> >> >> >
>> >> >> >  Anyone understand this?
>> >> >> >
>> >> >> > * Csound5GUI to work correctly (hopefully with FLTK) on all
>> >> >> > platforms]
>> >> >> >
>> >> >> >  I assume this is like one above?
>> >> >> >
>> >> >> >
>> >> >> >
>> >> >> > AND FEATURE REQUESTS
>> >> >> >
>> >> >> > * documentation for CsoundAC
>> >> >> >
>> >> >> >  I think Mike has been doing this
>> >> >> >
>> >> >> > * windows binaries distributed as a plain zip, without installer,
>> >> >> > and
>> >> >> >  without python dependency.
>> >> >> >
>> >> >> >  ??
>> >> >> >
>> >> >> > * Loris on windows
>> >> >> >
>> >> >> >  Does it not work?  My windows machine is currently too slow and 
>> >> >> > too
>> >> >> >  small-disked to do much trying
>> >> >> >
>> >> >> > * Multiple MIDI out ports
>> >> >> >
>> >> >> >  Hum
>> >> >> >
>> >> >> > * more PVS opcodes (formant extraction; shuffling/scattering 
>> >> >> > bins,
>> >> >> >  spectral tuning)
>> >> >> >
>> >> >> >  Sounds good when someone has time
>> >> >> >
>> >> >> > * Mine CDP opcodes/utilities, SAOL opcodes, Supercollider 
>> >> >> > opcodes,
>> >> >> > STK
>> >> >> > opcodes
>> >> >> >
>> >> >> >  Ditto
>> >> >> >
>> >> >> >
>> >> >> > * deferred tables for tablexkt opcode
>> >> >> >
>> >> >> >  The whole deferred table and power-of-2 needs revision
>> >> >> >
>> >> >> > * Beer for the developers
>> >> >> >
>> >> >> >  I liked this one, but Adnams is hard to get
>> >> >> >
>> >> >> > ==John ffitch
>> >> >> >
>> >> >> > -------------------------------------------------------------------------
>> >> >> > This SF.net email is sponsored by: Microsoft
>> >> >> > Defy all challenges. Microsoft(R) Visual Studio 2008.
>> >> >> > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>> >> >> > _______________________________________________
>> >> >> > Csound-devel mailing list
>> >> >> > Csound-devel@lists.sourceforge.net
>> >> >> > https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >> >> >
>> >> >>
>> >> >>
>> >> >
>> >>
>> >>
>> >> --------------------------------------------------------------------------------
>> >>
>> >>
>> >> > -------------------------------------------------------------------------
>> >> > This SF.net email is sponsored by: Microsoft
>> >> > Defy all challenges. Microsoft(R) Visual Studio 2008.
>> >> > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>> >>
>> >>
>> >> --------------------------------------------------------------------------------
>> >>
>> >>
>> >> > _______________________________________________
>> >> > Csound-devel mailing list
>> >> > Csound-devel@lists.sourceforge.net
>> >> > https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >> >
>> >>
>> >>
>> >> -------------------------------------------------------------------------
>> >> This SF.net email is sponsored by: Microsoft
>> >> Defy all challenges. Microsoft(R) Visual Studio 2008.
>> >> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>> >> _______________________________________________
>> >> Csound-devel mailing list
>> >> Csound-devel@lists.sourceforge.net
>> >> https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >>
>> >
>> > -------------------------------------------------------------------------
>> > This SF.net email is sponsored by: Microsoft
>> > Defy all challenges. Microsoft(R) Visual Studio 2008.
>> > http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>> > _______________________________________________
>> > Csound-devel mailing list
>> > Csound-devel@lists.sourceforge.net
>> > https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>> -------------------------------------------------------------------------
>> This SF.net email is sponsored by: Microsoft
>> Defy all challenges. Microsoft(R) Visual Studio 2008.
>> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel 


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft 
Defy all challenges. Microsoft(R) Visual Studio 2008. 
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net