Carbon latency when using hardware insert while tracking

[Burgher]

Does anyone know what the sample latency is for using a hardware insert while tracking in real time? Essentially how much additional sample latency would an additional DA/AD cause?

I have a Carbon but don’t have it setup yet as the studio build is ongoing. When I initially start recording, I was looking at using the Carbon mic Pres only, but then try integrating hardware inserts while tracking using specific gear to colour it or apply compression. For example, I like the idea of getting a drum mix and do some light parallel compression while tracking, but using an external hardware compressor on the drum stem, then send that back in to be distributed to the headphone mixes.

[Darryl Ramm]

I don't have a Carbon but am curious about this box.

Presumably the insert latency is <1ms, the same spec Avid provides for monitoring latency in Carbon.

You also likely want to know that when the track comes out of DSP mode and you use that compressor in a mix is the ADC compensation for that insert is correct. I would hope so, but given the amount of confusion around this stuff, and a few latency related bugs floating around I would want to see it actually work OK.

I would just measure things. A click track to a physical output. Physically split that signal to two inputs/tracks, one that then goes through a hardware insert using just straight through cable. And both those final tracks feed a monitor output, which you then record back to another track and look for a time shift between the two click signals you see on that track. I'd just use lots of going in and out pf physical I/O to guarantee Pro Tools thinks you are really monitoring this path.

Take stuff out of DSP and/or just play back the recorded tracks in the same setup and record that output see what happens.

[Burgher]

What you have stated is exactly what I would do to determine insert latency, and a good point about ADC with DSP off. Your method is exactly what I was planning too, but can’t do it right now.

Anyone else able to do this to confirm?

[eightamrock]

I ran this test a while ago. I saw almost no latency, the only latency introduced was whatever latency the hardware unit introduced itself. Im seeing sub 1ms latency its closer to .0005

[Darryl Ramm]

Neat

But am I just dense or shouldn't the track DSP button be on? And are you going out physical outputs and back into physical inputs... all to make sure to see real time latency, not say see a small error in ADC?

Measurements in samples would be great (and at what sample rate)... that would be interesting to compare to typical converter latencies.

[eightamrock]

Yeah so this was done with DSP OFF because I wanted to test specifically for using outboard gear during mixdown. I cant imagine with DSP ON it would be worse.

The method was using signal generator to create a test tone which I printed to the track and duplicated to a new track. I then put an analog insert on the new channel which routed through my patch bay and into an 1176 and back again through the patch bay and into the analog return. I then printed the track. So you can see the top is the original, the bottom is the copy with the analog insert printed onto it.

This project is 48k, you can see the sample count at the top. Its about 20 samples.

[Darryl Ramm]

Originally Posted by eightamrock

Yeah so this was done with DSP OFF because I wanted to test specifically for using outboard gear during mixdown. I cant imagine with DSP ON it would be worse.

The method was using signal generator to create a test tone which I printed to the track and duplicated to a new track. I then put an analog insert on the new channel which routed through my patch bay and into an 1176 and back again through the patch bay and into the analog return. I then printed the track. So you can see the top is the original, the bottom is the copy with the analog insert printed onto it.

This project is 48k, you can see the sample count at the top. Its about 20 samples.

You had ADC on correct?

My expectation with Carbon is that an hardware insert on a non-DSP track will just use the IO buffer, so in reality will depend on that buffer size (or 2 x the buffer size) and be significantly larger than what you see here... which I assume is the ADC corrected timing.

So I think what you measured is great for your needs... it is showing a small error in ADC correction for the insert. And you should be able to make this exactly zero using the ADC +/- correction on that track.

But it's not what the OP is asking about here with the real time latency of a hardware insert that you would see with monitoring (where ADC is not involved).

Somebody please throw something at me if I'm confused, again I don't own a Carbon. (feel free to throw a Carbon at me )

[eightamrock]

Originally Posted by Darryl Ramm

You had ADC on correct?

My expectation with Carbon is that an hardware insert on a non-DSP track will just use the IO buffer, so in reality will depend on that buffer size (or 2 x the buffer size) and be significantly larger than what you see here... which I assume is the ADC corrected timing.

So I think what you measured is great for your needs... it is showing a small error in ADC correction for the insert. And you should be able to make this exactly zero using the ADC +/- correction on that track.

But it's not what the OP is asking about here with the real time latency of a hardware insert that you would see with monitoring (where ADC is not involved).

Somebody please throw something at me if I'm confused, again I don't own a Carbon. (feel free to throw a Carbon at me )

OK so I am definitely willing to do this. What steps would I take?

I have monitored with analog insert on a channel (DSP ON) and not perceived any latency, but doesn't mean its not there.

[Darryl Ramm]

Thanks for for offering to do that.

Hopefully you won't perceive any latency.

Look at my comments above https://duc.avid.com/showpost.php?p=2680792&postcount=2 and try that.

Bonus points if you can measure at different sample rates.