Re: [Rosegarden-devel] more audio

[Richard B. <bo...@bo...>]

Today has been "ridicuously cute" day again.  Actually no it hasn't
really as the results are a bit patchy, however I've just implemented
waveform previews in SegmentItems for audio segment.  In other words
the Segments bear a little representation of the waveform they
represent.

http://www.masticate.com/rosegarden/rg-waveform.jpg

So far the results don't _quite_ look right as the changing resolution
and size of the SegmentItem pulls out some spikey harmonics from the
sample file.  It'd probably be more useful if we just used an RMS (root
mean square) value or similar approximation to the curve (bullshit
bullshit etc etc).

It'd probably also be nice to antialias the little representations
as well to make them look nice and smooth.  Although I think I'll
probably leave this up to Chris.

Newly recorded audio doesn't produce a valid preview at this time and
we don't yet do realtime waveform generation or even monitoring of
the audio input level - but these will surely follow soon - as hopefully
will some notion of "stereoness" and a basic audio level mixer.

I did actually manage to record an audio track yesterday whilst playing
several tracks of MIDI and some audio out - the timing is a bit hit and
miss at the moment as we can't yet compute end to end latency on the
JACK connections.  We'll either need to do some skullduggery to fix
it or hope that the JACK API soon sports these features in full.  I'm
certainly pushing for it.

B