Max Scheduler Lag (again)
Hi There -
I'm running Max 7 on a MacBook Pro w/ Retina.
I know there has been some discussion about this issue before, but I wasn't able to find much resolution - so I thought I'd test the waters.
I'm running a relatively low-complexity audio patch. I've replaced the metro object with a phasor~/edge~ arrangement, which helped a little bit. I'm opening Max in "low-resolution" and I've adjusted the vector size.
Still, I absolutely cannot get regular timing in the patch. There are very audible lags.
The main issue seems to be related to using the multislider as a "sequencer", as even the attached patch seems to work irregularly.
Is there anything else to try?
I'm not sure Low resolution helps at all. Are you sure you have Overdrive on and Scheduler in Audio Interrupt on?
Then, when it comes to your patch, I would suggest you use, say, a [table] instead of the [multislider] in your triggering path. Use a multislider to fill the table with desired value. I'm not sure how multislider behaves, it may make the signal switch to low priority thread.
Thank you, Jean-Francois.
I did have Overdrive and Audio Interrupt on - although, looking elsewhere on the board, it's suggested that they might impede on Multislider performance.
Table does work better - but I'm still confused as to why Multislider is so dysfunctional on Mac OSX.
While Table does work better, I'm still not sure it's enough of a solution. I know a lot of people use Max for more rhythmic music - am I the only person having timing problems here?
hi everyone-
still trying to rectify this - is there a better way to keep rhythmic sequencing precise in max?
this is what it sounds like, with the multislider.
Oops.
Yes - I have the identical scheduler settings. It's not so much that there is a difference between Metro and Phasor-based clocking, it's that both cases seem to be rhythmically imprecise, if you listen for awhile or analyze the output data in a quantizer.
Switching away from the Multislider does seem to help, but it is still a problem. I'm curious if there any solutions or if Cycling74 is aware of this issue and working on resolving it..
I'm 100% sure that multi slider is not your problem.
The timing in Max is sample accurate - but only if you stay in the signal domain. Your sequence clock uses edge~ which takes your timing from the signal domain into the scheduler which is not sample accurate. If you read the documentation then you'll find a more in depth explanation.
If you really need sample accuracy build a sequencer that never leaves the signal domain, it's not difficult to do.
if your scheduler is set to 1 or 2 milliseconds and overdrive is on - or if you use a signal driven solution - it is absolutely impossible that timing is not stable.
GUI objects wont influence the data flow either, if there is not enough room for data, GUI redrawing is the first thing which is skipped or lagging - metro is the last.
The timing of the above patch is very stable with Max 7.1 on my 2009 macbook pro. What is maybe confusing you is that the sound of the attack is sometimes sharp and sometimes round. That is correct behavior and happens because the envelope has no attack time and you start the sound with cycle~ at different (phase) moments (sometimes crossing zero and sometimes at its maximum or minimum value)