is it possible to set a buffer size in a m4l device?
hi,
i have some distortion-issues using a m4l device with a buffersize less than 512 samples. i actualy don't know why and couldn't find any help in this topic
https://cycling74.com/forums/problems-with-pfft-in-m4l-hearing-crackles-everywhere
but running ableton with this buffersize gives me some annoying latency when it comes to looping etc. so my question: is it possible to set a minimum buffer size for one m4l device? so only that device would have some latency which won't effect the other tracks. a could run this device with a 512 size buffer an 128 for ableton at global.
you would really help me out. thanks alot.
The buffer size in Live is a global parameter and thus can't be set for individual devices.
But regarding latency, be aware that each M4L device in an effect chain introduces latency proportional to the buffer size. So you could try to minimize the input/output latency by minimizing the number of devices in a chain, for example by combining several devices into one.
The MfL distortion issue occurs when Live's buffer size is not set to a multiple of 64 samples, and will be fixed for the next incremental release.
thanks for advices. but i have one m4l device in every track (ca.) and only one, containing a fft with large window, is doing these problems. i get distortion at 256 samples or less and not only in the track containing the m4l device but in everey track. iget even distortion in itunes or every other audio signal. so i guess even if the cpu-meter in ableton says 6% the cpu is having small peaks. the distorion is coming periodic.
but there must be a way for having a minimum buffer size in m4l, of course producing larger latency in this particular device, but that wouldnt be a problem for me. the device would have to buffer several vectors into one bigger buffer. it must be somehow possible.
and multiple of 64 didn't help, but e.g. 400 samples workes still fine.
any other ideas, or maybe tricks for the fft?
I'm having the same problem. While in editing mode everything runs smoothly. but going back into Live makes everything crack up. I can't see how it would be a problem having each MFL device have their own buffersize as long as it's a mulitple of the Live buffer size? Using the csound~ in Max works this way - if the csound buffer size is some multiple of the Max vector size everything runs just fine. I hope this issue will be fixed soon.