Sync to RC-505?

jamos

Heyell;

I'm trying to integrate an RC-505 looper with my dual G2 setup. teserday it worked great, today no so much.

The setup is: one G2 as master, running all the sequencers and doing audio; another G2 as synth, controlled from first G2, and the 505 running off the MIDI clock generated by the master G2, connected to the thru port on the second G2.

The looper is currently just playing a long unlooped vocal track that I recorded in sync with the G2. Again, it worked perfectly last night, but this morning the 505 is getting ahead of the G2s.

I suspect the problem may be the the 505 is losing sync due to the large mount of MIDI date sent from G2-1 to G2-2; but in that case I can't figure why it is getting ahead rather than falling behind. I will try using the 505 as the master clock, but that runs into some configuration issues as I need the second G2 to get data from the first as well as MIDI clock.

Anyone experienced with the 505? Any good resources for deep tech issues such as this?

jamos · Posted: Sat Mar 14, 2015 10:47 pm Post subject:

Well, I tried using the 505 as a master, no different. There's also a problem with the 505 not sending MIDI start/stop signals, as near as I can tell, so that's not a great solution.

However, weirdly enough, I decided it might be some sort of data corruption problem, so I rerecorded the track; it worked fine until I transferred the track to my PC via USB... and then the problem reoccured!

I will try again in the AM.

Tim Kleinert · Posted: Mon Mar 16, 2015 5:49 am Post subject:

I doubt that this is a MIDI clock issue.

Shoving a lot of MIDI data through one connection together with MIDI clock pulses normally causes the timing of the slaved unit to veer back and forth depending on load, but fall back in line when the load eases up.

Consistent drifting out of alignment due to MIDI would mean that clock pulses "get lost" in the data stream due to eg. underspec'ed stack buffer, but then you'd occasionally lose other stuff (notes, CC# messages) too. Either that or a poorly implemented clock pulse time-averaging/sync algorithm -both which I find unlikely.

jamos · Posted: Mon Mar 16, 2015 6:06 pm Post subject:

Thanks, Tim. i figured it out: operator error. I had both Tempo Lock and Loop Lock activated on the 505, neither of which are necessary in this application. Once those are turned off, everything is good.