Is it (still) the case that Pd's midi buffer is connected/synced with the audio buffer? (Sorry if any terminology incorrect).
Yes, MIDI messages are artificially delayed to match the additional latency caused by the ringbuffer (see "Delay" entry in audio settings resp. "-audiobuf" command line option).
At least in my scheduler-fix branch (https://github.com/pure-data/pure-data/pull/1756), this does not happen with "callbacks" enabled. However, this does not really help as you would typically need to use a larger hardware buffer size (confusingly called "blocksize") - which would again increase MIDI latency - and also increase MIDI jitter!
It would be great if the MIDI delay could be disabled globally, or even better - on a per-object or per-event basis! Would you mind opening a feature request on GitHub?
Christof
On 11.12.2023 13:34, Sam Ross wrote:
Hello list,
Is it (still) the case that Pd's midi buffer is connected/synced with the audio buffer? (Sorry if any terminology incorrect).
If this is the case, is it correct that the best way to get lowest latency control signals is to run control data in a separate instance that is removed from any audio processing, with the audio settings set for as low a latency as possible? So, delay 0 ms, I presume? And would altering the sample rate and block size also have an effect?
Thanks for any help you can give in assisting my understanding. Sam.
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> https://lists.puredata.info/listinfo/pd-list