Re: [Denemo-devel] midi input

[Jeremiah Benham]

On Wed, 2008-11-05 at 15:56 +0100, Nils Gey wrote:
> The idea of having immediatly playbacked the note is nice, however it should 
> be 
> the users choice to turn it off/on since I, for example, absolutly don't like 
> it. And if I need it I normally have my Midikeyboard connect to ZynAddSubFx 
> or Linuxsampler so I get sound anyway, without Denemo. Please note that with 
> JACK its no problem to route your midi-input to denemo and some other 
> programm at the same time.

It is already a switch in preferences->play back notes immediately. It
is currently broken at the moment. My desire was to create an interface
that I could test immediately.

> To the JACK-output. Do you mean JACK-audio output for the (CSound?-)playback 

csound already has jack audio output itself. There is no need for that. 

> or do you mean midi output?

midi output

> For the midi-output I agree that there has to be some amount of time to be 
> spent. Midi-output is not a easy thing and it should be well designed with 
> proper reactions to the Denemo-commands. Also there would be a massive amount 
> of new functions needed, for example tempo-changes/metronome-changes which 
> would be recognized by both Lilypond and midi, and universal-functions to 
> directly alter midi-commands, like the lilypond-entry function is now used 
> for lilypond. 
> 
> If Denemo should be able to output midi, which could be really amazing, then 
> it should be done properly. And if you ask me its a post 1.0 thing :)

Most everything has been written already in exportmidi. The only thing
is that exportmidi was written to output to a file. It also does one
track at a time. So it reads the data into a file line by line.
Exporting to jack midi would be mostly the same just reading the
vertically (multiple tracks simultaneous) instead of horizontally (one
track at a time). 

Jeremiah


> Nils
> 
> 
> On Wed, 05 Nov 2008 08:27:42 -0600
> Jeremiah Benham <address@hidden> wrote:
> 
> > On Wed, 2008-11-05 at 08:13 +0000, Richard Shann wrote:
> > > I have plugged in my m-audio keystation 49 and doing
> > > cat /dev/midi
> > > I can see the midi data coming in as I press the keys.
> > 
> > We can a universal function that responds to midi messages NOTE_ON,
> > NOTE_OFF, PITCH_BEND, etc... and if jack is not available we can default
> > to oss. Alsa does oss emulation so the user just has to provide the
> > device path /dev/snd/midi02, /dev/midi or whatever. 
> > 
> > > It looks like for the non-jack branch we could possibly just collect the
> > > simplest "noteon" data and use it as pitch input for Denemo. 
> > 
> > You are going to want to process it to determine if the input is
> > overlapping or if the foot pedal has been pressed. 
> > 
> > > This would
> > > still leave the question of how to get midi out, though. 
> > 
> > There is already some code written for this. Users who have a hardware
> > synth on their card can have immediate playback as notes are entered.
> > Unfortunately the code is broken at the moment. If
> > gui->si->prefs->immediateplayback is TRUE notes entered dnm_insertchord
> > calls playnotes() defined in midi.c. The call to playnotes it commented
> > out at the moment because it was broken. I am not sure how this works
> > exactly its using /dev/sequencer to write to. I believe it is then the
> > responsibility of your alsa drivers from there on. 
> > 
> > > You need some
> > > output as you play the music in as audible feedback that you have hit
> > > the right note.
> > 
> > Would we add jack output to the release also?  I could probably have it
> > sending out notes that are entered immediately. I could not do a full
> > score playback via jack without dedicating some time to it though. I
> > would want to get playnotes() working first anyway.  
> > 
> > Jeremiah
> > 
> > > It's not that I am avoiding getting jackmidi working on my computer,
> > > just a question of finding more than the odd half-hour to work on it.
> > > (You may have noticed that I keep doing lots of small things, for this
> > > reason).
> > > Richard
> > > 
> > > 
> > > 
> > > 
> > > _______________________________________________
> > > Denemo-devel mailing list
> > > address@hidden
> > > http://lists.gnu.org/mailman/listinfo/denemo-devel
> > 
> > 
> > 
> > _______________________________________________
> > Denemo-devel mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/denemo-devel
> 
>