Would i get any latency w/external comp for buss

[Detuned6]

I was wondering if this is the best way to add a external compressor(RNC) on some drum tracks.
What i am doing:
The digi 002 outputs(7&8)-->RNC L&R inputs - Rnc L&R outputs --> Digi 002 Inputs 3&4.

The outputs on the tracks channel strip (snare, trigger, kicks) are set to the channels 7&8. On a stereo AUX track, set the input to 3&4.

Will I be getting any latency? For example the snare track that is on the RNC aux buss, and the snare that is in the Stereo overheads tracks. Will there be any phase problems or any latency between the tracks?

Or, is there a better way to hook the rnc up ?

[daeron80]

Don't do it while tracking, because, yes, there is significant latency involved on a signal being recorded, unavoidably. But for mixing, if you need to compress a submix, the method you outlined works fine. Alternatively, you could submix to an Aux track and place an I/O insert it.

I don't know what you mean by "trigger." If it's for triggering MIDI samples or keying a dynamics plug, it doesn't need any processing. I would probably just put an I/O 7 on the kick track, and an I/O 8 on the snare track, then hook up the RNC ins and outs to 7&8 ins and outs. That would avoid the additional bussing latency. IIRC, the roundtrip latency for I/O inserts is 88 samples. It may or may not automatically compensate for that when playing from disk. If it doesn't, you'd need to either use Time Adjuster on all the other tracks, or slide K&S 88 samples early.

[daeron80]

Well, I apologize, I was completely wrong about that. If you assign an audio track (playing back audio from disk, not from an analog input) to an analog output, and bring it back in on another track through an analog input, PT automatically compensates, so the re-recorded sound has sample accurate placement (although certain plug-ins on the output track may or may not be compensated for). But if you use an analog I/O insert, you get a latency of

L = HWB * 2.34375 + 162

where L is total latency, and HWB is your hardware buffer setting. So, let's say your HWB is set to 128. 128 * 2.34375 = 300 + 162 = 462 samples of latency. Why that formula, I have no idea. I just ran the tests and analyzed them in Excel, and that's what it comes to. Analog I/O insert latencies for each HWB:

<font class="small">Code:<hr /><pre>
HWB Latency ms @ 48KHz
--------------------------------------
64 312 7
128 462 10
256 762 16
512 1362 28
1024 2562 53
</pre><hr />

So, the way you described is the best way to do it for mixing, although you could just as well bring them back on the same numbered inputs as the outputs used. But in order for PT to compensate, it has to be pulling the audio from disk. Audio originating at an analog input, sent to an analog output and back in again, will have enough latency to drive a drummer nuts.

[Detuned6]

Thanks for the info. My whole goal was to use my RNC on a buss w/pre recorded drum tracks(snare, snare trigger, kick trigger).I tryed differant ways trying to hook it up to the 002( like using the alt main outs,alt snc. inputs, using a buss,etc, but i get a phasey/flange noise and if i crank the rnc master vol I will get digital distortion) I dont want to record the compressed buss( becuase i cant change the comp settings once i am mixing) But i do want it present when i bounce to disk.
So since I am bring the RNC compressed tracks into PT on a Aux track, I am getting latency(?), but its dependent on the buffer setting(?) I can run the mix @ a 64 buffer, but once the vocals are done and i add more plugs, I will have to change the buffer to 512 or higher.
How do I compensate for the latency? If I record the rnc compressed mix on a stereo track, then I could nudge the track back the proper MS to fit in with the rest of the audio, but thats not what I want to do.
I tryed searching the DUC for basically a "how do i use a external comp for a 2 buss in real time". But i cant find much.

Thanks for the info daeron80, but i am still a little confused.

[NickH]

Hmm... Interesting...

I got the following figures using the good old null test:-

===================
(Low Latency Monitoring NOT on)
-------------------
762 Samples @ 256
1370 Samples @ 512
2586 Samples @ 1024
-------------------

----------------------------------

===================
(Low Latency Monitoring ON)
-------------------
87 Samples @ 256
87 Samples @ 512
87 Samples @ 1024

Cheers



N

[Detuned6]

<font class="small">Code:<hr /><pre>
HWB Latency ms @ 48KHz
--------------------------------------
64 312 7
128 462 10
256 762 16
512 1362 28
1024 2562 53
</pre><hr />

This only applies when i am recording correct? So when i am recording some vocals @ a 64 HWB, i would get 7ms of delay? If I am understanding this correctly.

[daeron80]

I got the following figures using the good old null test:-

===================
(Low Latency Monitoring NOT on)
-------------------
762 Samples @ 256
1370 Samples @ 512
2586 Samples @ 1024
-------------------

----------------------------------

===================
(Low Latency Monitoring ON)
-------------------
87 Samples @ 256
87 Samples @ 512
87 Samples @ 1024

Well, that's in the ballpark of the figures I got, also using null test. Hard to imagine what could have caused the slight discrepancies. I did the nulling by recording the signal to a separate track, then putting Time Adjuster flipped on the original. It didn't occur to me that Aux tracks might act a little different. Hmm... indeed.

BUT - Low Latency Monitoring mode applies ONLY to outputs 1 & 2, and inputs sent to 1 & 2 in LLM do NOT get included in a bounce. LLM mixes in a separate 20 X 2 mixer built into the 002 itself, where the 20 ins come from the 8 analog ins, the 10 digital ins, and outputs 1 & 2 of the PT software mixer. It's for tracking convenience only.

Something else to be aware of is that LLM has the same latency as a Hardware Buffer setting of 64. So HWB=64 with LLM is a waste. It's only helpful when you must use a higher HWB.

[daeron80]

This only applies when i am recording correct?

No, recording latencies are less than that. The list above applies only to analog I/O inserts. I'm going to be really basic here, hoping to be clear and not intending to insult your intelligence. When you click on an insert point, you see three choices: "no insert," "plug-in," and "i/o insert." The I/O insert menu shows you what input-output pairs are available in your I/O setup. If you choose one of those, it routes the audio through that output and back in that same numbered input automatically, for using a piece of external gear like a plug-in. I always assumed that was the best way to do it. But on an 002-based LE system, at least, it turns out that it's not necessarily the best, because it produces a ton more latency than using the track output to send to the gear, and returning to an Aux track.

[daeron80]

I dont want to record the compressed buss( becuase i cant change the comp settings once i am mixing) But i do want it present when i bounce to disk.

I can understand why. Once you get comfortable with your mixing skills, it's easier to get something you know will sound good, print it, and move on. But until you're sure it's right, it is nice to leave your options open.

So since I am bring the RNC compressed tracks into PT on a Aux track, I am getting latency(?)

Yes. I ran tests on that and discovered that latencies from an audio track with an analog output to an aux track with an analog input are:

<font class="small">Code:<hr /><pre>
Samples
HWB of latency ms
---------------------
64 306 6
128 458 10
256 762 16
512 1370 29
1024 2586 54
</pre><hr />

These are the same figures given by NickH. Now we know why the discrepancies - different method. These figures follow the formula HWB * 2.375 + 154. I'd love to know why. Seems like it ought to be HWB * 2 + 88 (two hardware buffers plus the A/D cycle). But it ain't. It's shocking how different the signal is after one pass through the 002 converters. Ack!

, but its dependent on the buffer setting(?) I can run the mix @ a 64 buffer, but once the vocals are done and i add more plugs, I will have to change the buffer to 512 or higher.

Yes, it is dependent on buffer setting, and it's usual to use lower settings for tracking and higher for mixing.

How do I compensate for the latency? If I record the rnc compressed mix on a stereo track, then I could nudge the track back the proper MS to fit in with the rest of the audio, but thats not what I want to do.

If you record it back in, PT will compensate and automatically put it in the right place for you - no nudging necessary.

If you mix using aux returns from the RNC, you will need to compensate. You have a choice of two methods for this.

1. Nudge the tracks that are sending out to the RNC; nudge them earlier by the number of samples in the chart above, according to your buffer setting.
2. Place a TimeAdjuster plug-in on all tracks except the ones that send to the RNC and the auxes it returns to, and set them all for the number of samples in the chart, according to the buffer setting.

Method 2 is standard, because when you start scooting tracks around, it's really easy to forget to make careful notes, and get confused about what's moved and what's not, and by how much.

Don't worry about being confused. As this thread shows, I was pretty confused about it myself, and full of all sorts of false assumptions about it (left over from years of TDM use) until I ran the tests. If you need any more help or clarification, please don't hesitate to ask.

[daeron80]

So when i am recording some vocals @ a 64 HWB, i would get 7ms of delay? If I am understanding this correctly.

For charts of recording latencies at each sample rate and buffer setting, see my web page:

http://www.elvenminstrel.com/recordi...002latency.htm