Host tempo/bar changes....

bbulkow

I ran into a problem, think I figured it out, and if so it's a reasonable operational mode (not a bug), but might require some explaining to not have it confuse other people?

I was running into a problem where mobius wasn't in sync with ableton as master. I had set the (new groovy session settings) as directed in a thread I initiated in Feb where the first track is host-sync'ed on a bar basis, and the subsequent tracks are driven by the first track with loop granularity. v41 .

I think what was happening was the plugin initializes, picks up at some point a start point and a speed. If the host starts and stops - even with no tracks recorded - I suspect Mobius is keeping the old timings. This happens at startup, as far as I can tell.

I guess I expect that if the host starts and stops or changes tempo with no tracks recorded, then mobius will pick up the new values.

The answer appears to be always global reset after starting the host's time the way you like it. Global Reset will certainly pick up the new timing.

Is there some setting about when to pick up new time, or how to deal with time changes? I understand that host timing changes could be handled a lot of ways - stretch, reset start, etc. That seems like a hard problem and maybe there are a lot of config settings I haven't found / don't understand.

At first I'd like the simple thing - if no tracks, always go with new.

If my diagnosis is correct, I would request:

Pick up new timing if there are no tracks recorded (including the "sync track from master").
If "sync track with master" enabled, and the host has changed and the track has not resync'ed in any way, a visual indication that track is not in sync would be groovy.

Right now, I hope / believe I have a workaround, and it's sensible. Whenever I touch the tempo or start/stop, do a global reset after not before (I think I had been doing a GR, then changing the tempo or start time, then attempting to loop).

Thanks again for a great program / plugin! The new "session" settings are FAR more discoverable!

bbulkow

Hm, I'm not sure I'm to the bottom of this yet. While I was mostly getting the right tempo and bar start, one time (low key performance) I seemed to get the "out of sync" again (loop start def. not on bar, maybe on beat, maybe not). I decided to focus on making loops that didn't depend as much on exactly when they started and ended because I could see visually that the loop length was precise and correct. I also got into an odd situation where the first loop's start and end was different from the second and third, that was using the technique of having the first loop sync to host, and the subsequent loops sync to track. I suspect this was "pilot error" because I likely re-recorded the first loop while still having loops in the 2nd and third, so I would imagine that first loop could get a measure off.

I still like the idea of some visual indicator that the start-end is not lined up (to host or among tracks) for those of us still beginning and not slopping around our loops as much (maybe soon!)

I'll be spending time over the coming weeks attempting to fully characterize.... look forward to any hints on the matter.... how is it supposed to work in these corner cases..... it's certainly true in the "heat of the moment" one can get a little more confused than at home....

alexs1

Hm, not sure what it could be. I know for me in FL in Windows the tempo changes on host synced tracks respond with the corresponding values before, during, and after changes in the Host’s BPM without having to flush it with a GlobalReset. There might be a weird quirk in how Ableton reports pulses, or how plugins handle host sync information in Ableton. Lots of users have it rigged to Ableton though, there might be someone who’s configured it similarly and could reply. In the meantime, maybe look into Ableton’s manual if there’s anything that relates?

bbulkow

Thanks, I see that ableton 12 and FL support both VST2 and VST3 plugins, there shouldn't be any sheninigans related to something as simple as a wrong or outdated version or protocol handling.

I just did a little research, and AI says there is a hot discussion on the JUCE forums about a "meta parameter bug", as well as an older discussion about "timeInSamples" and tempo changes - and that FL had some of the same "Quirks" (around timeInSamples).

I think my job here will be to try to narrow down a very specific case, now that I know the intent is immediate update in the face of tempo and start-stop events.