Subject: Re: [linux-audio-dev] Needing advice on implementing MIDI in my app (RTMix)
From: Paul Davis (pbd_AT_op.net)
Date: Fri Oct 11 2002 - 01:44:42 EEST
>1) First off, is there a simple GPL library that is easy to implement
>and that focuses on MIDI? If not, any website with some sample code
>would be greatly appreciated!
libmidi++, used in ardour but entirely separate from it, does
this. however, its written in C++, and i suspect you will find it
quite confusing since it also makes much use of libsigc++. you may
find useful code there however, since it has a full MIDI parser (based
on the one from keykit) including MMC and MTC.
>2) I am trying to design something that is not driver specific (i.e. OSS
>or ALSA), so is there any way to do this (i.e. sending message directly
>to /dev/midi device or something similar without blocking it to other
>apps)
yes, but not without blocking it to other apps. you can't do this
without being "driver specific", since the non-blocking aspect
requires a design that moves it away from being "raw midi".
>3) For checking the presence of the midi interface on a Linux box, is it
>safest to look for /dev/midi device, or is there a "nicer" way?
that will work (for OSS and ALSA).
>4) Is there a way to send MIDI data in such a way that it is perceived
>by other app on the system (i.e. like sending a midi signal through
>midi-out which is via cable connected to midi-in on the same machine,
>except without doing so physically)? If so, how? If not, what are the
>alternatives (other than above-mentioned physical solution)?
you can just use a name pipe (FIFO) in the exact same way you use
/dev/midiNN. libmidi++ has support for this.
>5) Finally, this may be a bit far-fetched question, but here it goes
>anyhow: does anyone know if Pd and other applications like it are able
>to receive simple numeric messages via network (TCP/IP) protocol (since
>this would be an alternative to #4)?
i had a recollection that someone had implemented OSC for Pd, but i
may be wrong. OSC would be a much better solution if only a decent
reference implementation of OSC existed.
--p
This archive was generated by hypermail 2b28 : Fri Oct 11 2002 - 10:33:47 EEST