[sdiy] MIDI Clock sync advice

brianw brianw at audiobanshee.com
Mon Mar 11 02:25:30 CET 2024

One thing I forgot to mention is that time stamps allow transport jitter to be corrected. So long as the time stamps themselves are not subject to jitter, and the largest jitter delay is known, then the sequencer can force a fixed latency between input and output. The accurate time stamps (if available) can be used to schedule output MIDI at a fixed latency / delay. That latency could be 5 ms, 10 ms, or maybe as high as 12 ms (depending upon other latency in the signal path), but the musicians will automatically adapt to reasonable latency after the jitter is corrected / removed.

This is where technology like CoreMIDI and MOTU's MTP are beneficial. They're not available everywhere, though.


On Mar 10, 2024, at 6:19 PM, brian wrote:
> On Mar 10, 2024, at 6:08 AM, Mike Bryant wrote:
>> High speed USB is 125uS, but in general the cheaper MCUs used in most musical instruments still work on lower speed.  But in any case it's still prone to an element of jitter which may or may not be a problem.  I'd love to find a real drummer accurate to 1mS :-)
> Accurate and precise are two different aspects of timing measurement. You can't just dismiss one or the other arbitrarily. That's true for all scientific measurements.
> When it comes to music, we also have the human ability to correct for latency, but not for jitter. So, an accurate drummer might be ahead or behind the other musicians, based on where their kit is set up on stage, but that timing error is constant if the drummer has a good "feel" - the other musicians all subconsciously adapt to the latency.
> Acceptable latency is 10 ms to 12 ms in the studio. A band playing together on a stage experiences delays of 5 ms to 10 ms due to the speed of sound. But the feel of a drum beat requires timing down to some smaller level to preserve the feel. If the snare is 5 ms behind, it sounds in the 'groove.' 7 ms ahead and it has 'snap.' But if the hits are moving around by a significant fraction of that, due to jitter, the 'feel' will not be consistent. In an October 1987 article from Electronic Musician, Craig Anderton quotes Larry Fast as saying that controlling the sound with precision under a millisecond makes synthetic drum patterns sound more realistic (human).
> That's not an exact figure - less than a millisecond - but it says to me that we should probably aim for an order of magnitude better than what people can hear, just to preserve the feel.

More information about the Synth-diy mailing list