Page 1 of 1

Posted: 19 Sep 2011, 21:47
by GDLive
Hello !!

I have been trying for two weekends to have a patch working.

Whatever I do, I have some crackles using this patch. I post here a simplified version (audio in is continuous oscillator, fade in and fade out to avoid beginning and final audio clic ).

With this version, I'm joining an audio excerpt of this typical problems.

Rec + background record + transpose (>0)

The CPU jauge is no more than 5%
I have this problem :
- on my laptop (win 7 64 bits) + fireface UC => whatever Asio buffer I set
- on my laptop (win 7 64 bits) + alternative driver (directsound...)
- on my workstation(win XP sp3 32 bits) + hoontech DSP 2000, both asio and directsound drivers

I have this problem in 44,1 kHz, as in 96 kHz..

Asio driver don't see any error/dropout during the crackle.

With same settings, I don't have troubles with other patch, it's seems to be with the sampler background record + rec + play + transpose setting. I don't understand why. If I don't use the background record option, the result is ok, then I don't think it's the transposition fonction only that produce the problem.

I someone could reproduce this pb or tell me what is wrong, he would be very welcome !! :D

Here is the patch (try to do other thing with this patch playing, moving windows, selecting other tracks,...)
http://www.sensomusic.com/forums/uploads.php?file=ForumHelpRequestPatch.pat

And an audio example of my problem follows.

Thanks,

GD.

Posted: 19 Sep 2011, 21:51
by GDLive
And the sound extract :

http://www.sensomusic.com/forums/uploads.php?file=crouc_1-track-3.zip

(wav.zip)

edit : precisions about the link

Posted: 21 Sep 2011, 18:38
by GDLive
OK, I received an explanation by the boss himself :
The problem is the result off my bad comprehension of the circular buffer technology.
Even if the playing time is smaller than the recording buffer size, using circular buffer could not give the answer.
In fact, I've to create/pilote my own buffer (with buffer reset at the right moment) using 3 or more sampler in my case (shift >0) or following a grain approach.

Posted: 21 Sep 2011, 19:18
by senso
thanks to answer +++