MIDI Clock Unstable on MIDIO128

[m00dawg]

My MIDIO128 seems to not want to keep a stable BPM over USB or the MIDI ports. Using Ableton Live in slave mode shows that BPM being woefully inconsistent. Likewise, using my GameBoy (via LSDJ and my custom ArduinoBoy) causes music to play 2-4x faster. This is true when using my own MIDI files (generated from Reaper), a blank MIDI file recorded on the MIDIO128 and a random MIDI file on the Interwebz (this one specifically).

Looking at MIOS Studio and MIDI Monitor show otherwise normal data, although I do see an awful lot of clock updates (not sure if that is normal or not). MIOS Studio thinks the BPM is the same as what MIDIO128 shows though neither Live nor LSDJ show anything close. Live's BPM drifts from 100-300+ constantly and never settles down like it normally does when in slave mode off an external clock.

I'm running the bare essentials at the moment - just the standard CS with a 2x20 LCD and an SD card. No IIC_MIDI or anything like that. I thought maybe the CS might be causing grief so I disconnected it while playing a MIDI file but to no avail.

Anyone have any thoughts about how I can go about troubleshooting this?

[m00dawg]

Aha figured out a bit more. If I start MIDIO128 up and immediately put it the BPM on LOCK, I can sync the GameBoy at the set BPM. Live still flips out some of the time, but if I start Live fresh and start the song in MIDIO128 and then hit EXT, it will lock on, but only until I stop the song. If I try to play a song on MIDIO128 again, Live still looses it.

Also, if I have the BPM set to Auto or MSTR and then go to LOCK, the displayed BPM isn't correct, though I can change the BPM up and down from whatever the MIDI file was set to. Also of note, I think MIDIO128 is having trouble reading the clock data from the Reaper exported file. Looks like it's playing at around 300BPM instead of 105BPM. Reaper and Live can read the tempo data without issue though.

MIDIO128 plays the test song I grabbed off the Interwebz at around 150BPM, though Ableton and Reaper can't seem to find a master clock in that file.

[TK.]

There is something special with your .mid files, could you please attach one to this posting, so that I can check why the clock rate isn't determined correctly? This would be the first step to analyze the issue.

 

Best Regards, Thorsten.

[m00dawg]

Thanks for looking into it TK! I put a ZIP together of basically everything off the SD card.

Celebration.mid is the file that came from Reaper. It starts at 105BPM and goes to 80BPM at measure 82 and then does a slide down to 67 starting at measure 88.3, then starts a slower slide up to 105 starting at measure 89. There is only a single track - a 4 on the floor style thing at various points we used with MBFM (sammichFM). I have been focusing on the clock issue so I haven't tested that out yet, but just wanted to let you know it's there.

REC1.MID and REC2.MID were blank recordings I made off MIDIO128. The first only went for about a measure, the second for a minute or so I think.

daft_punk-get_lucky_ft_pharrell_williams.mid is the file that came from the Interwebz which, I think, might be at 150BPM. Neither Reaper or Live found any clock data but 150BPM is what I was able to see from the MIDIO128 when playing the track when I was able to get Live to behave and give me an accurate BPM reading.

Hopefully that gives you something to go on!

MIDIO128_BPM_Test.zip

[TK.]

Thanks for the files!

 

This was a simple bug: the MIDI clock was output at 384 ppqn (inherited from the original MIDIbox SEQ V4 code), but a .mid file can have it's own ppqn rate which was ignored.

 

It's fixed now, please try this version: http://www.ucapps.de/mios32/midio128_v3_017_pre1.zip

 

Please note that the REC1.MID file won't output a proper clock whenever the song is looped, because re-loading from SD Card takes some time at which no MIDI clock event is sent.

 

Best Regards, Thorsten.

[m00dawg]

Yep that seems to work flawlessly with the intended BPM behavior. I have discovered what may be a bug in Reaper but as far as MIDIO128 goes, that change seems to have done the trick. Thanks TK!