Best way to measure total I/o latency?

[Ru_C]

Hi all, I'm just trying out Pro Tools Ultimate with 28day trial, & wanted to find the best way test total I/o latency, compared to my TDM rig.

I can obviously read the theoretical specs online, & of course just *listen*, but I'd like to know 'scientifically' what I'm dealing with .

Rig#1 is a TDM system running PT 10 HD
Rig#2 is a Pro Tools Ultimate 2020 software with a Firewire interface (yes I'm testing this on an old 5.1 Mac pro). I'd prefer to to it with a Native Pcie card & connect it to the same pair of AVID I/O units, but that's all I have for now.

Should I be physically patching an output back to an input, playing (for instance) a snare sample out, then looking at the gap between the two, in order to measure total round trip?

is there a better way?

Cheers

[falcowe]

Maybe RLT Utility will help automate the process a bit?
https://oblique-audio.com/rtl-utility.php

[albee1952]

Not the "empirical" answer, but maybe the important answer;

First off, using a firewire (or USB) interface is going to be nowhere near as good as the TDM system and HD/Native(with HD IO or 192 IO) is going to perform significantly better than firewire, but not as good as the TDM rig. BUT***, what's really important is, will it be good enough for your needs?

My own experience with firewire(003) and HD/Native(in the same machine):
With the 003 and a 64 buffer and no latency-inducing plugins, I rarely heard any complaint about latency in the headphones, even with drummers playing my Roland TD10 kit triggering Addictive Drums. With the Native card and HD IO(actually, 2xHD IO and 2x192 IO), still the 64 buffer setting and keeping plugin latency down to nearly none(actually, 11 samples), I never heard any mention of latency in the headphones.

This was always 48K and 24 bit(higher sample rates will have lower latency given the same buffer setting). If you wish to track with UA plugins, then you would NEED to use Apollo hardware and the UA mixer window. On an HDN system, UA plugins have too much latency to track with(I used to have them all inactive while tracking, but after a year or two, I just dropped UA). Hopefully, that gives you the info you needed, but if you really want exact numbers, play a track that is routed out an output, connected to an input and recorded to a separate(muted) audio track(muted so you don't create a feedback loop). Then use Tab to Transient(show samples or time in the counter) and measure the difference

[Ru_C]

Originally Posted by albee1952

... but if you really want exact numbers, play a track that is routed out an output, connected to an input and recorded to a separate(muted) audio track(muted so you don't create a feedback loop). Then use Tab to Transient(show samples or time in the counter) and measure the difference

..thank you for answering the thread...I felt like an idiot for posting after I figured it out !

I plugged a guitar in the channel one input, & sent it to channel 2 input via a pre fader send (at unity) as I recorded ..then zoomed in to measure the difference in recorded waveforms.

I'm using a pair of stock AVID 16 I/O's for combined 32 I/o, & the latency I measured was exactly what has been published in the past by AVID it seems.

44.1khz Native sessions measured around 3.5ms (maybe 3.3 ms)...but, 96k native sessions were about 1.8ms, which is actually a tiny be better than the 1.9ms I measured using TDM at 44.1khz...