003R Round Trip Latency with PT9

[AndyJCP]

Dear DUCers,

despite low buffer settings (64 samples at 44kHz), I still have some issues with latency. It's ok for tracking guitar and even drums, but I still feel it's a bit uncomfortable for tracking vocals.
So, first I want to understand and that's why I have a few questions for you more qualified guys and girls.

First Question: How is the round-trip latency generated? My understanding:
1) Input latency of my analog gear - Should be near zero, no?
2) Analog to Digital Conversion of the 003R - How much is that?
3) Buffer-induced delay - should be 1.4ms in my case (64 samples at 44kHz)
3a) ??? Auto delay compensation ??? - does this induce any latency at all? I thought it just shifts the tracks to be in line but does not induce any additional latency?
4) Digital to Analog Conversion of the 003R - How much is that?
5) Output latency of my analog gear - Should be near zero, no?

Is this list complete?
Can you give me figures for 2) and 4)?

Many thanks in advance,

Andy

[albee1952]

1) Input latency of my analog gear - Yes, zero latency
2) Analog to Digital Conversion of the 003R - How much is that? Around .5 to .8 ms
3) Buffer-induced delay - should be 1.4ms in my case (64 samples at 44kHz)I'll accept that(too tired to do the math)
3a) ??? Auto delay compensation ??? - does this induce any latency at all? I thought it just shifts the tracks to be in line but does not induce any additional latency? It works by delaying everything to the track with the most delay. IOW, if you insert a plugin on any track, that has 1000 samples of latency, all other tracks will be given latency to match. This is why you should not use high-latency plugins during tracking.
4) Digital to Analog Conversion of the 003R - How much is that? Same is going in A>D-also dependent on sample rate
5) Output latency of my analog gear - Should be near zero, no? yes

Read 3a again as that is the crux of this issue(I suspect). Nothing can make audio happen earlier, so ADC simply backs things up. Take note of the latency of the plugins you routinely use. If you see some/any with latency over 10 samples, then reserve those for after all tracking is done. Alternately, you could turn Delay Compensation off while tracking.

[AndyJCP]

Hi Albee,

thanks for the good answer!
I don't have any latency-inducing plugins during tracking, so I can forget about 3a).
This would sum up to 0.8 + 1.4 + 0.8ms which gives 3ms only.
Which is similar to standing 1m away from my loudspeakers...

So it should not be noticable.
Still, I hear a strange offset in my head when singing.
Maybe it's some kind of strange phasing going on inside my head?!?

Does anyone of you have an issue when singing with SW monitoring at 64samples buffer size?

Andy

[nst7]

I've never had an issue with singing through a 64 buffer. It sounds like real time to me, no phase issues, etc. Same with guitar through amp sims. And I do this on a 2007 Imac with plugins on various tracks, reverb on vocals, etc.

I would check to see if something else is going on that would cause this.

[AndyJCP]

Hi nst7,

yes, I was also suspecting some other problem in my chain.
I don't believe 3ms can cause such an effect. I will search for another possible cause of the latency.

Andy

[AndyJCP]

I have made some further tests.
I still perceive latency when recording my voice at 64samples buffer.
It's much better when using the Low Latency Monitoring on my 003.

I think (but I'm not sure) I have the problem when I use a low volume on my headphones. Then I hear my voice in the recording room and in the headphones, which makes a strange effect...
If I turn up the volume, it gets better.

Anyone else noticed this? Or do I only have to get better isolated headphones?

[getz76]

Originally Posted by AndyJCP

Anyone else noticed this? Or do I only have to get better isolated headphones?

Isolation headphones won't help much with vocals, as the vibrations of the singer is still happening through the skull. You are better off muting the vocal track being recorded or adding a bit of reverb to your monitor.

Any reason you are not using 32 buffers with the 003R? It should be available.

[daeron80]

Here's what I measured with my 002R. I don't know where the additional delay comes from, but it is there. Still, 7ms is nothing. It really should be practically inaudible. Like hearing yourself through an analog system with a loudspeaker 2 meters away. That's exactly the way I perceive it. If it's more delayed than that for you, there must be something causing extra delay. You can test for yourself and see. Maybe you're just sensitive, though.

[AndyJCP]

Originally Posted by getz76

Isolation headphones won't help much with vocals, as the vibrations of the singer is still happening through the skull. You are better off muting the vocal track being recorded or adding a bit of reverb to your monitor.

Any reason you are not using 32 buffers with the 003R? It should be available.

I'm on PC, so I can only go down to 64.

[AndyJCP]

Originally Posted by daeron80

Here's what I measured with my 002R. I don't know where the additional delay comes from, but it is there. Still, 7ms is nothing. It really should be practically inaudible. Like hearing yourself through an analog system with a loudspeaker 2 meters away. That's exactly the way I perceive it. If it's more delayed than that for you, there must be something causing extra delay. You can test for yourself and see. Maybe you're just sensitive, though.

That's very good data, daeron! And I was also looking for a methodology to measure the latency. I will do the test and post the result