some MIDI clock testing results


Firstly I’m not sure if here is the right place to be writing, so let me know if I should join the dev list and post results there.

I’ve only done a quick check tonight without debugging turned on (I will recompile and try again probably later)…

What I have noticed is the slave is basically working, When I press play the system starts to play, when I press stop the system stops. The responsiveness is very good, I’m running this at 48khz with a period of 64 samples and it is nice and smooth.

However here are some things I noticed that could be improved to make it more complete.

1.) When I press stop on my external sequencer the transport moves back to the start. This means that even if I press “continue” then the transport starts as if I have pressed “Start”.

2.) When moving location in a sequence on the sequencer the transport in ardour doesn’t move (this doesn’t happen when moving the transport or , so at this stage the sequencer can’t be used as a proper transport control including locations. (this would be the proper way for it to operate in slave so all MIDI and audio time is kept constant.

3,) This is only cosmetic, but ideally the display when set to Bars and Beats should show the true position of the track as defined by the incoming MIDI clock, or should be disabled (as a workaround though). I know other sequencers have implemented this and it was a very good way to test whether there is drift from the clock.

That’s it for now. I’m recompiling with the debug flag turned on and I’ll post some more results over the weekend. (Please let me know if these results should go to the user or dev mailing list instead.



I have repeated this with debug enabled and wrote the output to a log file, as per Hans instructions (with trace MIDI input turned on just so the messages from the sequencer can be seen… there are a lot of MIDI clock messages showing).

Where should I send this file?