On Sat, May 26, 2018 at 12:10:03AM +0200, Sebastian Reitenbach wrote: > Am Freitag, Mai 25, 2018 22:22 CEST, Alexandre Ratchov <a...@caoua.org> > schrieb: > > > On Fri, May 25, 2018 at 10:22:26AM +0200, Sebastian Reitenbach wrote: > > > Hi Alexandre, > > > > > > Am Freitag, Mai 25, 2018 08:41 CEST, Alexandre Ratchov <a...@caoua.org> > > > schrieb: > > > > > > > On Fri, May 25, 2018 at 12:21:18AM +0200, Alexandre Ratchov wrote: > > > > > > > > > > Maybe, but unlikely. If the device attaches, generally it works. Show > > > > > what's received on midi2 when you type on the keyboard or do any other > > > > > simple actions. > > > > > > > > > > > > > fwiw, here's a small utility that I use very often to debug & connect > > > > MIDI ports. > > > > > > > > http://caoua.org/alex/obsd/midicat.tar.gz > > > > > > > > To connect two rmidi/2 -> rmidi/1 and dump the data on stderr, run: > > > > > > > > midicat -d -q rmidi/2 -q rmidi/1 > > > > > > Actually, that midicat seems to do the trick. When I run it in debug, > > > and press some keys on the > > > system-1, output looks like: > > > > > > f8 > > > f8 > > > fe > > > f8 > > > f8 > > > f8 > > > 90 80 9b > > > > ^^^^^^^^^^^ > > > > these 3 bytes make no sense. The status byte 0x90 is correct but the > > two args are wrong. This is when you pressed the key, right? > > > > > ... > > > 90 9b b2 > > > > These are also wrong, and this is when you released the key, right? > > Interestingly the 0x9b is repeated. > > > > f8 > f8 > f8 > 90 80 34 > 90 40 f8 > ^^^^ released key
This one is cleary wrong. 0x90 must be followed by two numbers in the 0..0x7f range. > > > > Clearly, the midi events are corrupted by the usb interface. Does the > > midi interface claim to be "class compliant"? does it come with > > drivers for MacOS or Windows? > > It was advertized as class compliant, and it just came the cable, no drivers. > But as I said, it was el-chepo from ebay not even 3 EUR incl. shipping from > China. > do you know if it works on other operating systems?
