Re: [Denemo-devel] Denemo broken ....

[Richard Shann]

On Wed, 2011-08-03 at 20:48 +0200, R. Mattes wrote:
> On Wed, 03 Aug 2011 18:14:01 +0100, Richard Shann wrote
> > On Wed, 2011-08-03 at 18:52 +0200, R. Mattes wrote:
> 
> > > But why does jack_output_midi_event take the velocity from the
> > > movement and ignores whatever velocity gets passed in? Do we
> really
> > > never want to play back a crescendo?
> > This probably an instance of Denemo's confusing names, this is
> might 
> > be some command for making a sound effect or some such. The midi
> generated
> > by Denemo controls the velocity in the expected sort of way so it
> > wouldn't be going through this (at least in the non-jack code). 
> 
> ??? But that is the code that shows up in the stacktrace. So, at least
> change_tuning does call it. Now, what will be sent to jackd? The first
> two bytes of the tuning change message, than one byte of "velocity"
> and
> nothing more ....  broke beyond repair. 

As I say, the jack code probably *is* broke (perhaps beyond repair), and
is currently being replaced by Dominic Sacre`
On fluidsynth the tuning message is not sent, because the synth is NULL
at this point in the code. In that code the use of the master volume for
the velocity is restricted to NOTEON messages, and again as I say, it is
presumably not used by the general MIDI playback mechanism, as otherwise
you couldn't make a crescendo. As this code is currently being replaced
it would be worthwhile avoiding it at present, if you can.

Richard