Latency

[Rob]

My interface is about 15 years old already, but it still works fine. I've been eyeing an RME UCX for a while but I can't justify it.

[legobeats]

Beeing a long time MOTU user I switched to a M4 last year and can recommend it. On my Mac Mini I can go down to 32 samples without any glitches. Usually I play with 64 to have some performance overhead.

[perceval]

I must say I like Steinberg because of their background. But I have zero experience with them, so don't take this as dogma!

MOTU had a rough spot some years ago, but they have been turning things around, and I only hear/read good things about their newer gear, on the pro side, and even the hifi side.

But I will forever stay away from any Presonus product.

[mprinz]

My interface is about 15 years old already, but it still works fine. I've been eyeing an RME UCX for a while but I can't justify it.

perceval and Rob you are right, I kind of hope that.

What is the latency of the first class drum modules?

I mean, isn't it the same signal "path"?

[mprinz]

Hi,

I bought a Steinberg UR22C USB3 and tested it yesterday.
My M-Audio Fast Track Ultra (I found that it was presented at NAMM 2008) driver showed a latency
of 3.1 ms and the Steinberg driver in its lowest buffer 1.9 ms.

I returned it.
My ears were saying ... oh... it feels way better, and my brain said: Hi dude, it's ridiculous thinking you can here a difference of 1.2 ms.

It was actually the "fast" setting. There three options, I remember something like "fast", "standard" , "stable" .
Switching to the "standard" lead to values even nearer to 3.1 ms, so, I saved 130 €.

I mean, maybe the sound was actually better? Better DAC? Difficult so say.

Regards

[mprinz]

Hi guys,

what is the simplest way to measure the external latency?
A have a small oscilloscope here.
I would connect th 1st to the microphone (or the piezo output (pad)) and the 2nd channel to
the headphone cable. Something like that?

Greetings

M. Prinz

[Rob]

This is what I have done in the past.

mesureLatency.jpg (173.09 KiB) Viewed 6992 times

[mprinz]

Ok,

so you assume the input latency (input jack -> DAW timeline peak) is zero?
I don'nt know it. But I was told: Noooo, don't do it like so. You have to use an independent system.

I used an oscilloscope and measured the time from the first head piezo signal change to the first headphone transient.
I got something like 11 ms with the Millenium 10" snare pad head piezo, a Millenium MPS-850 module, USB MIDI,
M-Audio Fast Track Ultra audio interface (2008), and the internal HP-amp.
I didn't check it with an eDrumIn module.

When do the scan start? When it's going to negative voltage or after the drop when it rises above zero?
Have a look to the picture, please - 8ms or 11ms ? Anyway in real life this will not make a difference ...

IMG_20200113_230226_1.jpg (895.94 KiB) Viewed 6988 times

I was told, those Millenium/Alesis/Fame/Xdrum have a way higher latency than Roland modules.
I don't know.
So assuming the Millenium/Alesis MIDI creation time or their "there was an event - play a sample" time for internal sounds
is not higher than 0.1 ms, as-well, they will have set a way higher scan time, that is not adjustable?
Or can a DA converter be that slow? I read something of 1ms for the DA conversion.
Or is it just a fairy tale of Roland lovers?

I also turn off the "local control", meaning that the module is only sending MIDI data and not playing it's samples, in order to "save" time
... Does that make sense?
Would the next scan really take "more" time, if the module is playing a sound ? It is SET to a specified value ...

The fact that an acoustic snare sound has a 3 ms latency until it reaches the ear at 1m distance is often forgotten.
The bass drum correspondingly more ...

[Rob]

so you assume the input latency (input jack -> DAW timeline peak) is zero?
I don'nt know it. But I was told: Noooo, don't do it like so. You have to use an independent system.

No, I do not. But because both recorded tracks (the microphone input, and the input from the headphone out) have the same input latency, they cancel each other out.

You're method should also work fine.

When do the scan start? When it's going to negative voltage or after the drop when it rises above zero?

What would your ears say? I think you ears would perceive the sound starting at the first significant peek regardless of weather it's positive or negative. To me it looks to be about 10ms in the pic you provided, but remember that measurement is dependent on your audio interface settings because it includes the audio output latency.

[mprinz]

Hi Rob,

ah yuuur. I haven't read the hole sentence and than it was already clear.

The image: I was asking with regard to the piezo transient not the headphone signal.
When do you start your scan?