juliochristo

The note on event was just an afterthought but did not tie in directly with the clock offset question. The lag from a PC audio buffer IS well into the ms range, this is the lag I would want to try and compensate. But after thinking about this problem again, it occurred to me that the clockbox probably can't do any kind of offset compensation because it has no reference to offset from. It is the audio that has to offset in comparison to the incoming clock tick.

I'm looking at the code and trying to determine where the best place to implement a delay in order to add drift compensation to the midi clock. The use case is when using this clock with the PC, the PC will have lag before the audio hits the sound card but my synth will not. So the clock should have some sort of compensation for this in order for all the sound sources to be in time correct? The most obvious place I found was in Tick(void) just after MCLOCK_Tick(); Thoughts? As a side note: looking through the source code, if this clock method is the same method used by the MidiBox SEQ projects, there is no clock prioritization scheme I could see. So if a clock and a note on event where to occur at the same time, would the clock have to wait for the note on and therefore introduce clock drift?

I *could* do it on vero (I just hate working with the stuff) but I would still need a programmed mc as I do not have a programmer so I would still have to order *something* at the very least.

Thanks for the replies. For those of you who have built this before, how long does SmashTv usually take to ship? I have a gig coming up on May 16th that I need to use this for and I'm curious how long it will take to receive it and if I can build it fast enough (I'm fairly experienced with an iron and have all the equipment) to actually use at the show. Tom ps. Now what would be ideal, is combining this project with the midi merge project. I'm using a MOTU 4x6 merge matrix (which does MTC but not midi clock...which I could never understand that choice) for that purpose but ideally the less hardware I have to tote around the better.

Has anyone done any testing with the clock generated for accuracy? I'm considering building this project as I need a rock solid midi clock but I want to be sure this will be stable enough for what I am doing. If no one has, would someone mind posting a dump from midi-ox of the generated clock timing (or similar data)?