Notifications
Clear all
Topic starter
In another discussion ( https://www.yamahasynth.com/ask-a-question/interesting-signal-chain ) I brought up how the timing on triggered arps has jitter and latency.
In a few tests, I found that arps can have a varying latency of 13-23ms with respect to other non-arp parts triggered simultaneously. This is uneffected by tempo or arp Sync Quantize settings
My first example used Cubase, but I found an even better way to demonstrate this with just the Montage:
1. I loaded a drum performance in Part 1 and copied it to Part 2. All FX and Arp SFX are off.
2. Part 1 has an arp with a simple 4/4 kick
3. Part 1 is panned hard left
4. Part 2 has the arp off
3. Part 2 is panned hard right
(I have attached the single X7B user performance to this post)
In this recording, The first 8 seconds are me playing key C1. Part 1 on the left triggers the arp and Part 2 on the right plays the same kick (no arp)
http://sneak-thief.com/demos/Arp-NoArp_vs_NoArp-NoArp.wav
Result 1: Kicks on the left side from Part 1 (the triggered arp) come 13-23ms later than the kicks on Part 2 (right)
Result 2: The second 9 seconds are with both arps off. In the audio file, drums from Part 2 (right) consistently comes 1ms later - which seems to be normal given Yamaha's common practice of giving higher parts higher priority.
Posted : 04/08/2019 9:45 am
May be interesting to look at the click output along side these waveforms to get an idea where the internal tempo clock is relative to these observations. You may have run out of channels to correlate (for high correlation) since L and R are already used - but you may be able to use a 4-channel audio interface and connect main L&R to 2 channels and assign L&R to the other 2 (and send click out assign L&R).
You could do other experiments to see if there is any adjustment necessary by using the Assign path vs. Main path. This would be fairly dead easy if Click Output could be set to Main L (or Main R) mono -- like allowed for AssignL vs AssignR. Maybe one of the click types is mono -- I haven't run through the 10 types.
Perhaps you've already looked at this since you say things related to clock do not seem to matter. But I still think it would be interesting to see how the clock relates to these events you're documenting.
Posted : 05/08/2019 4:42 am
