Hello everyone!
   I have problems using the pvs opcodes. I tried this sequence:
    fstart pvsanal ain, 32768, 8192, 32768, 0
    aout pvsadsyn fstart, 16384, 1.0

   This works quite OK, but if I set the kfmod parameter of pvsadsyn to >1 I 
get really horrible delays. And I tried and tried, but didn't find parameter 
constellations, that could resolve that.
   Can any one help?
   Kindest regards
           Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

I suspect it is because you are using  pvsadsyn (oscillator bank) to
re-synthesize 16384 bins. Why not try with pvsynth (overlap-add) or
reduce the number of bins?

Best
Peiman

2009/3/6 Julien Claassen :
> Hello everyone!
>  I have problems using the pvs opcodes. I tried this sequence:
>   fstart pvsanal ain, 32768, 8192, 32768, 0
>   aout pvsadsyn fstart, 16384, 1.0
>
>  This works quite OK, but if I set the kfmod parameter of pvsadsyn to >1 I
> get really horrible delays. And I tried and tried, but didn't find parameter
> constellations, that could resolve that.
>  Can any one help?
>  Kindest regards
>          Julien
>
> --------
> Music was my first love and it will be my last (John Miles)
>
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>

Thanks. But isn't there another way. I'm now working with other pvs-opcodes as 
well and I gather, the smaller the number of bins, the worth is the quality.
   Is there perhaps a correction function or some other fancy way of 
eliminating delay while having best possible quality. I don't mind realtime. 
Main thing is output quality.
   Kindest regards
         Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

Try using pvsynth instead which uses an overlap-add re-synthesis.

Best
Peiman

2009/3/6 Julien Claassen :
> Thanks. But isn't there another way. I'm now working with other pvs-opcodes
> as well and I gather, the smaller the number of bins, the worth is the
> quality.
>  Is there perhaps a correction function or some other fancy way of
> eliminating delay while having best possible quality. I don't mind realtime.
> Main thing is output quality.
>  Kindest regards
>        Julien
>
> --------
> Music was my first love and it will be my last (John Miles)
>
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>

Peiman!
   I'm using pvsynth already. But still the problems keep on. I'm working with 
pitchscaling using pvscale and I've tried others, and it's very easy to 
reintroduce the delay and other artifacts. They sound very undesireable, 
especially, when my ears tell me, that the "clean" signal hidden in there 
sounds very good!
   Any ideas what I can do about that? Please.
   Kindest regard
          Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

Ah I thought you mentioned pvsadsyn in your message!

Can you post your csd here?

2009/3/6 Julien Claassen :
> Peiman!
>  I'm using pvsynth already. But still the problems keep on. I'm working with
> pitchscaling using pvscale and I've tried others, and it's very easy to
> reintroduce the delay and other artifacts. They sound very undesireable,
> especially, when my ears tell me, that the "clean" signal hidden in there
> sounds very good!
>  Any ideas what I can do about that? Please.
>  Kindest regard
>         Julien
>
> --------
> Music was my first love and it will be my last (John Miles)
>
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>

Hello!
   Yes in the original example I mentioned it. But here's the relevant code. 
It's rather simple beyond that.

[orc header]
instr 1
   a1 soundin "new.wav"
   fstart pvsanal ain, 2048, 128, 2048, 0 ; get it started
   kfr, kamp pvspitch fstart, 0.06 ; for future experiments
   fpitch pvscale fstart, 1.4, 2, 6 ; pitch it
   aout pvsynth fpitch ; resynth it
   outs aout, aout
endin

   That's it, upto now. I'm working on extending it. But the output now sounds 
stupid. You can recreate it by recording a tiny bit of spoken word. Or I could 
upload my file.
   Kindest regards
        Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

The only issue I can see in your code is that ain is not declared
anywhere. But I'm sure that's not the issue.

I have no problems with the csd bellow. Change the soundfile for one
of yours. Are you on mac or windows? And what sample k-rate are you
using? Strange.... unless you are running out of CPU?







sr = 44100
ksmps = 10
nchnls = 2

instr 1
 a1 soundin "new.wav"
 fstart pvsanal a1, 2048, 2048/4, 2048*2, 0 ; get it started
 kfr, kamp pvspitch fstart, 0.06 ; for future experiments
 fpitch pvscale fstart, 1.4, 2, 6 ; pitch it
 aout pvsynth fpitch ; resynth it
 outs aout, aout
endin




i 1 0 10
e







2009/3/6 Julien Claassen :
> Hello!
>  Yes in the original example I mentioned it. But here's the relevant code.
> It's rather simple beyond that.
>
> [orc header]
> instr 1
>  a1 soundin "new.wav"
>  fstart pvsanal ain, 2048, 128, 2048, 0 ; get it started
>  kfr, kamp pvspitch fstart, 0.06 ; for future experiments
>  fpitch pvscale fstart, 1.4, 2, 6 ; pitch it
>  aout pvsynth fpitch ; resynth it
>  outs aout, aout
> endin
>
>  That's it, upto now. I'm working on extending it. But the output now sounds
> stupid. You can recreate it by recording a tiny bit of spoken word. Or I
> could upload my file.
>  Kindest regards
>       Julien
>
> --------
> Music was my first love and it will be my last (John Miles)
>
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>

Hi!
   that was just a typo with the audio input. I'm running on LINUX and my 
samplerate is 48000 kr = 4800
   I'll try the exact values you used for pvsanal and recheck if the output 
sounds better.
   Thanks so far!
   Best,
          Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

Hi again!
   I tried and it still sounds very underwater.
   I just uploaded the complete thing here:
http://juliencoder.de/v.tar.bz2
   It contains:
test.orc and test.sco - csound input files
new.wav - sound-source
out.wav - output I get with that setup
   I already went on tweaking to imrove it. Do you have suggestions to achieve 
a better result?
   Kindest regards
        Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

Hi,

The ulr is not working for me. Just attach them to the message maybe.

good luck
P

2009/3/6 Julien Claassen :
> Hi again!
>  I tried and it still sounds very underwater.
>  I just uploaded the complete thing here:
> http://juliencoder.de/v.tar.bz2
>  It contains:
> test.orc and test.sco - csound input files
> new.wav - sound-source
> out.wav - output I get with that setup
>  I already went on tweaking to imrove it. Do you have suggestions to achieve
> a better result?
>  Kindest regards
>       Julien
>
> --------
> Music was my first love and it will be my last (John Miles)
>
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>

Hi!
   OK, I attached it. I don't understand why the URL doesn't work. Tried it 
myself and failed. Strange...
   Can you have a look and listen?
   Kindest regards
          Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

Did you mark the file as publicly readable?


There does seem to be some clear differences between pvshift with 
formant correction and without; Victor will be best placed to answer 
that. In the general case, pvsadsyn has some limitations, principal 
amongst which is that it is not good with large frame overlaps (in terms 
of samples). This is because only bin amplitudes are interpolated 
between frames. It was written on a 333MHz machine, where efficiency 
seemed by far the most important concern. Large overlaps are not all 
that good even with FFT resynthesis. But with a small overlap as you 
have it in your example, it works  reasonably well with pitch shifting, 
though not sadly with formant correction. With a sample rate of 48000, 
an FFT size of 2048 should do very well, and perhaps up to 4096 if you 
want to resolve very low frequencies (as for drums). Certainly there 
will be no advantage in using really large windows of 32768.


Richard Dobson

Julien Claassen wrote:
> Hi!
>   OK, I attached it. I don't understand why the URL doesn't work. Tried 
> it myself and failed. Strange...
>   Can you have a look and listen?
>   Kindest regards
>          Julien
> 
> --------
> Music was my first love and it will be my last (John Miles)
> 
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe 
> csound"

sorry I've been very busy today. OK it sounds fine here on osx. Does
it happen if you write to disk? Or only in realtime? Are you running
the latest version of csound? Single or double?


2009/3/6 Julien Claassen :
> Hi!
>  OK, I attached it. I don't understand why the URL doesn't work. Tried it
> myself and failed. Strange...
>  Can you have a look and listen?
>  Kindest regards
>         Julien
>
> --------
> Music was my first love and it will be my last (John Miles)
>
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"

Hello!
   been away myself.
   I use:
Csound version 5.08.91 beta (double samples) May 26 2008
   on linux with write to disk.
   Did youcompare your output with my old output?
   Kindest regards
           Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

Hello Richard!
   Thanks for the explanation.
   as to the download, there seems to be some server trouble. I'll have to do 
some more checks. Nothing to do with readability at all, I'm afraid.
   I think first of all, I need to get the newest csound. But I always fear 
installing it a bit, because of the loris and ATS stuff and all that. :-)
   Kindest regards
         Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

ahh I just did,

your output sounds OK to me! at second 0.5 there is a clip (output
amplitude is to high) but apart from that I don't hear anything
unusual.

Peiman

2009/3/7 Julien Claassen :
> Hello!
>  been away myself.
>  I use:
> Csound version 5.08.91 beta (double samples) May 26 2008
>  on linux with write to disk.
>  Did youcompare your output with my old output?
>  Kindest regards
>          Julien
>
> --------
> Music was my first love and it will be my last (John Miles)
>
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>

Hello Peiman!
   It sounds to synthetic and underwatery to me. Do you think that I could 
achieve better, more natural sounds with praat?
   Kindest regards
        Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

Oh, such a pity. I've seen, that praat seems to be GUI only. that leaves me 
out. So more experimenting with csound... I guess.
   Or does anyone else have an idea, which other tool to use, which other 
csound opcodes to try?
   Kindest regards
           Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

Hi Julien,

Yes it does sound unnatural, but then it is difficult to create
natural vocal transpositions (bordering on impossible in my
experience). A good tool I have come across is the new version of
audiosculpt which has a very convincing result for pitch shifting +
keeping formant (you can control the percentage at which the formant
is kept or discarded). I have never used praat or LPC so I can't
comment there (except that I gave up LPC with csound due to bugs a
while ago! I hope it's fixed now). Of course AS is mac only and it's
not free :-(

Best
Peiman



2009/3/7 Julien Claassen :
> Hello Peiman!
>  It sounds to synthetic and underwatery to me. Do you think that I could
> achieve better, more natural sounds with praat?
>  Kindest regards
>       Julien
>
> --------
> Music was my first love and it will be my last (John Miles)
>
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>

Peiman!
   Could you run my voice sample through your software anyway, I'd like to 
hear. Say with a pitch factor of 1.4 to 1.5, something like that, with good 
formant keeping. Would you do that and mail it or upload somewhere? I'm 
curious!
   Kindest regards
         Julien

--------
Music was my first love and it will be my last (John Miles)

======== FIND MY WEB-PROJECT AT: ========
http://ltsb.sourceforge.net
the Linux TextBased Studio guide
======= AND MY PERSONAL PAGES AT: =======
http://www.juliencoder.de

ahh sorry I just replied to your message in the other thread by mistake.

Best
Peiman

2009/3/7 Julien Claassen :
> Peiman!
>  Could you run my voice sample through your software anyway, I'd like to
> hear. Say with a pitch factor of 1.4 to 1.5, something like that, with good
> formant keeping. Would you do that and mail it or upload somewhere? I'm
> curious!
>  Kindest regards
>        Julien
>
> --------
> Music was my first love and it will be my last (John Miles)
>
> ======== FIND MY WEB-PROJECT AT: ========
> http://ltsb.sourceforge.net
> the Linux TextBased Studio guide
> ======= AND MY PERSONAL PAGES AT: =======
> http://www.juliencoder.de
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>