Scarlett 18i20/18i18 latency in AUM

Timespare

Is there anyone on this forum that has a Scarlett 18i20/18i18 or any other 3G Scarlett for that matter paired up with the 2020 iPad pro 12.9”?

I’m curious to see your Round Trip Latency in AUM with 64 frames buffer size.
I currently have the Steinberg UR44 with these values:

I’ve heard that the Scarlett 3G should have really good RTL so I wonder if this would be worth the upgrade for me. I also need the extra two channels ofc 😊

In my current setup I use Synthstrom deluge as a looper so I bus send the audio from any given channel to the deluge and it works well except for the latency that the round trip produces. If I route the audio directly from the source to the deluge the recorded audio is on point.

Janosax

Unfortunately those AUM values are unreliable, they’re always almost the same with any interface.

You will need Latency Meter app: https://apps.apple.com/fr/app/round-trip-latency-meter/id1427507645

It’s the easiest way. Here with Scarlett Solo Gen3:

Here with iRig Pro IO, huge difference:

Or you will need to measure it manually with something like any kick drum sound in AUM recorded with a loop back cable output to input on your interface then measure latency in AudioShare. Here 17 ms RTL with iRig PRO IO at 64 buffers.

Sound quality on this Scarlett interface is really good, and low latency has helped my sax playing timing too:

https://forum.audiob.us/discussion/46449/brheaking-lines-ipad-pro-m1-sax-music-making-dawless-experiment-helium-multitrack-recorder#latest

Timespare

Hi @Janosax! Thanks for your reply. Very helpful.

Here is the result from UR44

So almost the same as the Scarlett.

Although the manual measure tells a different story.. 50ms?

Janosax

@Timespare was it an audio file or a virtual instrument sound? Any limiter used? This can adds latency. Try with an audio file loaded in AUM file player. Not sure if that could change anything, but here Latency comp for hardware inputs is off:

Timespare

@Janosax Ok, I did the test again with your suggested method and I got this result:

So it's better but twice as much as the measure from LatencyMeter. 17ms is like your iRig and that's not really workable.

Janosax

Here Scarlett value is same 8 ms:

I initially made a Ruismaker Noir kick drum one bar recording, played in file player sent to output then input is recorded (with physical cable plugged between output/input):

Timespare

@Janosax Thanks! I redid it exactly like that and now I got:

So I guess there will not be any improvement at all in latency with the Scarlett over the UR44.

Another issue is that 64 frames buffer size is not really usable if you use a lot of AUv3 synths and effects in AUM so you have to bump it up to 128 or 256 and then the RTL is sky high with any audiointerface.

Janosax

My iPad Pro M1 works quite well at 64 buffers, I need to use lot of AU before needing to switch to 128. I’ve used several years iRig PRO IO at 128 buffers on 7 Plus, which is 20 ms RTL, and was really used to that. It’s doable but depends on your instrument, how you feel latency, and how you adapt to it. Now using 8 ms 64 buffers RTL, and can certainly feel the improvement for my ryhtmic playing, my recordings are more tight. 10 ms 128 buffers is still ok for me, and 2 X less 20 ms I was used to. Using good iems like Shure SE215 helps a lot for isolation and to compensate timing.

espiegel123

Btw, 20 ms may sound on paper like a lot of latency, but it is the same latency as standing 20 feet from your speaker.

Timespare

@Janosax That's interesting to hear that there could be a reason to upgrade to the iPad pro M1 even though a lot of people says that it's not a big jump in performance over the 2020 iPad pro.

@espiegel123 Yeah I know but I'm extremely sensible so I'll hear it immediately if something is of by 20ms in the mix.

Have a listen to these two clips and see if you can hear where the baseline is off beat.