I discovered this when troubleshooting non-functional sync between Fostex D2424LV multitrack and Pyramid.
The setup is simply Fostex -> laptop USB -> Pyramid, the Fostex as the midi clock master (it can only slave to MTC and the like) with a midi monitor running on the laptop to see what’s going on. Because the delay between start and clock is so ridiculously big (five seconds), it’s really easy to see with your plain eyes when hitting play on the Fostex:
0s Fostex sends Start
-> Pyramid starts playback immediately according to its own bpm/clock
… time passes…
5s Fostex sends first Clock
-> Pyramid clearly picks up the incoming bmp/clock at this point
-> …but by now things are hopelessly out of sync
This is really easy for me to reproduce, so since you say it’s not supposed to behave like this, I’ll test with v2.30 this evening to see if it’s just a regression in the 2.9x beta and report back.
FWIW, I also tried with direct midi-cable connection instead of USB, it behaves the same that way. I also tested this with a keyboard with a simple internal sequencer in place of the Pyramid in the above setup, and the keyboard does wait that 5s before starting playback.
Thanks for looking into this!