Another thread got me interested in trying PipeWire + JACK again as the last time I tried it (last year) I would get freezing during export. Since then I’ve been happily using ALSA. When I tried PW+J again today I got the same freeze, but for the few ranges that did export I noticed it was much faster than using ALSA, so I’d like to try and get it working.
I’m using 512 with ALSA and I tried various settings with PW+JACK. I do occasionally get the same issue when using ALSA so perhaps it is a system configuration issue and I’m just seeing it more regularly with JACK.
Wow, that are a a lot of ranges. So this likely a different issue (as opposed to pipewire getting stuck on freewheeling switches).
Before export, could you open the Log Window, and see if there is an error message?
“TransportFSM::bad_transition …” or “Duplicate Transport State Event …” or “Cannot prepare transport for export” - Those have meanwhile been fixed for multi-range export.
Yes, it might be worth a shot. Export has significantly changed (now there’s also MIDI stem export etc). – be sure to create a backup of the session before opening it in A7.
If the issue persists, could you share a session that causes it, or provide a step by step recipe how to reproduce the issue?
My test with A7 seemed to be going well. But I got to about 270 ranges and then received this error (both with JACK/PW and ALSA):
I noticed 1 (possibly 2) other bugs with the nightly build. Do I report these or do I wait until A7 is released before reporting?
Is it a higher or lower buffer that makes the export faster? And is there a reason Ardour doesn’t use the fastest settings possible automatically when exporting?
Looks like I already made the feature request a couple of years ago and forgot about it.
As mentioned in the pre-release warning, IRC is best to discuss issues.
The feature request is a good one. Although since the list has individual checkboxes it is not trivial to allow range-selection toggles. It’ll be a bit of work to make this happen.
“TransportStateChange” events are not handled while exporting, and just accumulate in the event-queue.
Those events are not relevant during export and can be ignored, but it looks like 2 are added for each export range (start, stop transport). After 512 such events have accumulated … boom.