I was curious to know, what was the actual latency with external input of Octatrack, as I could here by ear that something is wrong, so I did a test :
I plugged-in Rytm mk 2 main output to octa A B and send a metronome. Then I routed the main output of Octa to Ableton. Octatrack is a master clock.
I activated Octa’s metronome as well and routed it through cue out to second track in Ableton.
I recorded 2 tracks and figured out that I experience 5ms of latency ( I presume "Duration 00:00:005’’ is 5ms ), which is way too much.
I’ve run the same test with delay compensation on / off in Octa.
The same issue with Digitone.
Anyone else with the same latency? Am I having an issue with my device or I may missing out something?
Delay compensation on the OT delays the DIR output so that a signal listened to through a thru machine does have the same latency as thru DIR. If don’t listen via DIR, the setting won’t do anything. 5ms isn’t much, though. OT needs time to process. Lots of DSP operations introduce latencies. You can compensate in your DAW.
Rytm is plugged to Octa thru track. Then 2 metronomes (of Rytm and Octa) are sent to 2 separate tracks in abletons through main out and cue out.
That’s how I can measure an input latency.
I’ve tried measuring another way : I simply routed rytm metronom to thru machine, recorded a metronome in Octa. the latency is around 140 - 160 samples, instead of 64 as stated in the manual.
5ms can be too much when syncing stuff, where a frame of reference is a factor, definitely very noticeable, 2-3ms is less noticeable but still apparent, material dependent. The OT and Rytm both have pretty tight midi clock without too much jitter, so an audio delay on the OT thru can be heard if Rytm is being sent into OT thru and heard from its own output simultaneously.
However I believe the OT thru latency to be in the 2-3ms range, so it would only really be a problem if it can be heard (material dependent) or when wanting to capture precise loops. Of course there are ways to compensate using micro timing etc so in many cases it can be worked around.
To the OP it might be worth double checking the setup and results as 5ms seems too much, any fx on OT thru?
Honestly at this point whenever I am recording hardware I am always committing to the practice of editing the microtiming. Those small amounts of ms can really **** up the transients. I’ve accepted that this is just a part of modern production even with some dedicated timing device like the ERM Multiclock (I have a USAMO myself).
Of course a counter argument to this could be that many of the records we grew up listening to didn’t worry about these hyperedit details. This may be mainly a modern era discussion. One might even say that those sync gaps (or even jitter) contribute to various machines having a unique “groove.”
Btw, there’s a nice free plugin from DDMF, makes meassuring much more easy. Just use the mouse to highlight an area in your DAW and that little thing provides meassurements in bars, time down to ms and in samples. The reading in Live’s arrangement only provides time down to milliseconds.
for ex: if I’m layering it in a DAW to something locked onto the grid (like manually placed audio samples of drums). especially with kicks
also for my 2c: even if we don’t hear something outright our plugins/DSP will react accordingly. compressors with certain settings will react to peaks that we cant even hear reliably but result from the interaction of different signals
like I said though- this is some hyperfocus, so YMMV. there’s nothing wrong with choosing to say “**** it” and just make music
I know it’s an endless discussion, but allow me a quick thought regarding jitter and “unique groove”. Even if you don’t take into account the technical side (the 808 for example is actually pretty tight. If I’m not mistaken, in the circuit is one trigger that is distributed to all voices. Basically that trigger just gets multed and muted/unmuted for each individual voice accordingly when you set a step for a voice. Pretty simple, reliable mechanism. It’s a pretty simple, straight-forward sequencer, it doesn’t have to generate individual triggers for all voices which might lead to timing issues.), the “unique groove” thing is, as I understand it a myth. Jitter are basically random timing deviations, it can’t inject any sort of groove into your material, because the timing jumps all over the place (on a micro scale), randomly. Think how when you would randomly distract a drummer. You’d just mess up the groove.
Thus there is good reason in trying to minimize jitter. Latency on the other hand can easily be dealt with, because it’s a fixed timing deviation. I guess we’d all intuitively alter the material in order to compensate anyway, i.e. forming the attack portion of a sound, or using microtiming.
Or set up everything in the DAW to compensate for different latencies.
