Triple oscillator not working from rc2.9. #3292
Comments
bieltoni
changed the title
|
The problem occurs when you play the Triple Oscillator within the song editor. It seems as if buffers information gets corrupted. |
|
Can you provide a video, please? |
|
Theres definitely something wrong. I can reproduce it when TripleOSC playes with kicker together and TripleOSC has a OSC with sawwave. Then the sound stops and the peak of 3OSC goes to max. I guess it's something with: #3145 |
|
@BaraMGB Can you provide a small project that can be used to reproduce the problem? Thanks! |
|
What OS are this relevant for? If any track is muted in the problem project then, Couldent this be the 'silenced B&B-track' bug? |
|
@BaraMGB Here is another test file. A video where it works correctly And a video with the latest version of LMMS with the bug. I hope it helps. |
|
I have tested with https://lmms.io/lsp/?action=show&file=10220 and also with "Surrender-Main" from the "shorties" files and here are some more findings:
|
…ync") (#3145)" This reverts commit e3e14bb in an attempt to fix issue #3292. The problem that's addressed is very elusive: * It cannot be reproduced reliably. * It seems to be more prone to appear in release builds. * It only seems to appear when two or more instruments are mixed into one channel. * If you solo a problematic Triple Osc it goes away. The observations above seem to indicate that there's rather a problem with the mixing (due to multithreading?) and that the changes made to Triple Osc seem to trigger this deeper problem more quickly.
|
I have reverted the changes from #3145 with commit cf26da3. However, due to the symptoms described above I think it's rather a problem with the mixing and multithreading. The fact that soloing a Triple Oscillator makes the problem go away and yields correct sound also seems to indicate that there is nothing wrong with the computation of the audio and the update of the phase information. But as we are in an RC phase right now I guess it's better to revert the changes. @bieltoni Can you please check if the bug is solved once there is a newer RC available? Please only close this ticket if you find that it's solved. Thanks! |
|
For me the issue seems fixed. |
|
It would be a big setback, if the off-phase behaviour is re-instated. |
|
@michaelgregorius I have check in a new system (build in Debian 9 without KXStudio sources) and the last git version, works perfect. |
…ync") (LMMS#3145)" This reverts commit 291942f in an attempt to fix issue LMMS#3292. The problem that's addressed is very elusive: * It cannot be reproduced reliably. * It seems to be more prone to appear in release builds. * It only seems to appear when two or more instruments are mixed into one channel. * If you solo a problematic Triple Osc it goes away. The observations above seem to indicate that there's rather a problem with the mixing (due to multithreading?) and that the changes made to Triple Osc seem to trigger this deeper problem more quickly.
When a track with a tripleoscillator is played, a sound distorsion occurs. I have compiled different versions of lmms to determine from which the problem occurs and the rc2.8 version works correctly. From the version rc2.9 the tripleoscilator stops working correctly.
Debian 8 x64 with KXStudio repositories.
The text was updated successfully, but these errors were encountered: