[Rosegarden-devel] audio_etc branch report

[Chris C. <ca...@al...>]

Well, it might be worth taking a look at it again now.  The basics are 
pretty much in place: you should be able to play a goodly quantity of 
audio tracks with large or small files and repeating segments, with 
plugins, and with level monitoring at the GUI and basic one-file 
recording.

Things that work in HEAD but are not currently enabled or working on 
the branch include asynchronous audio playback (the Preview button in 
the audio segment manager), audio in loops, audio monitoring of 
recording, adjustment of buffer sizes (they're both hard-coded and 
not particularly optimal at the moment) and probably a few other 
things.  Also it will probably crash on exit.  It shouldn't crash 
during use though, except possibly if JACK isn't running (I haven't 
tested that greatly).

Other things to be addressed include the fact that although audio 
playback is (supposed to be) sample synchronous, recording is not 
yet; and although playback is supposed to be correctly synchronous 
with MIDI, that depends on having them both sync'd to the same PCM, 
which the ALSA queue is probably not by default.

I've only tested this with up to about 30 tracks of audio so far 
(there are still only 16 hardcoded audio instruments but of course 
you can have more than one track per instrument) consisting of a 
mixture of short repeating beep.wav files and longer recorded files, 
with the odd plugin.  Running as a non-privileged user on a stock 
SuSE 9.0 kernel I have no problem running 30 tracks without dropouts, 
but we could do with a bit more testing there.

So, the branch is called three_thoughts_on_audio_etc, give it a try.


Chris