Latency when using inserts / Are inserts considered a channel?

Two questions:
For example, if I connect an external (hardware) compressor, do I also have the 0.7 ms latency? I’m thinking about using an external, high-quality compressor for a drum bus. Do I have to expect additional latency? At least the DA/AD conversion of the channel will take some time. But is that the 0.7 ms or less?

2nd Question:
Do the physical inputs and outputs used for this insert count as channels, i.e. are they subtracted from the maximum usable 48 channels?

The .7ms latency time is the theoretical maximum latency for audio passing through the entire console - excluding any external routing. It is going to be the same for all audio passing through the system. (This is assuming the inputs are all coming from the same place - either all locally on the back of the console, or all from a stagebox, etc. Different input sources (local vs stage box, etc) will actually have a different total system latency).

Effectively, this means it is a “wash” for all audio passing through the console. You do NOT need to consider the .7ms system latency when worrying about external routing. You DO want to measure the added latency that any external routing adds however and potentially factor that into your calculations.

The easiest way to measure this is to set your system up without the external routing and get a measurement (using OpenSoundMeter, SMAART, etc). This will be close to .7ms, but it won’t be exactly .7ms which is why you need to actually measure it. Then add in your external routing and take another measurement. The difference between the two measurements is the added latency that your external routing adds and it’s this number that you potentially need to worry about.

EDIT - actually the initial measurement might be higher than .7ms depending on the method used to capture the data. For example, if you use the USB interface in/out of the console to a computer running your measuring software, it is going to measure higher than .7ms. However this “baseline” time doesn’t matter at all. (It could be 5 seconds long and it isn’t going to matter). That is because all we care about is the difference between the baseline measurement without the insert routing engaged and the measurement with the insert routing engaged.

[Back to the original post]
That being said, you really only need to worry/compensate for the latency that this external routing adds if you have identical audio paths (ie the same channel) going through the system with one going out the external path and one/more NOT going out the external path. In that case, you will want to delay the audio paths that are not going out the external routing by the added latency amount.

If ALL of your drums are going through the external system, and there are no drum channels that fail to go through the external device, then there is usually little need to worry about the added latency of the external routing of your drum buss. (I wouldn’t worry about it unless the extra latency was really high for some reason - like 12ms or greater).

However, if you are using this external drum processing like a “Drum Crush” group and you ALSO have your drum channels going through a drum buss without this processing and you expect to mix the regular and crushed audio together, then you absolutely need to add the external latency number to the regular drum buss that isn’t routed externally.

Second Question - If you are really using “insert points”, then the external audio does not count as channels (ie one of the 48 available channels). However they do need to be routed out/in to the console which you will do either through analog or digital I/O. They will count towards the I/O counts. For example, if you want to use Dante to move this audio out and back into the console, you will need a pair of Dante Outputs and Inputs (assuming your drum buss is stereo) available to actually send this audio through. The same hold true for other digital routing (like USB) or even analog audio routing (ie you’ll need enough physical analog XLR or 1/4" plugs to accommodate this external routing).

If on the other hand you are sending channels/busses out of the console and then route a return back into the console as a new channel/buss, then yes that WOULD count towards your 48 available channels. It’s a similar difference between using FX racks as inserts or with FX busses. Using inserts doesn’t count as a busses, but the buss routing does.

1 Like

I also once looked for the additional latency when using an analog insert, but couldn’t find anything.
I then simply tested it acoustically (not measured) myself with Pink Noise and came up with an unexpectedly relatively high value of approximately 0.31 ms when using a local output and input.
Perhaps someone can confirm this – preferably A&H. )

You can also test it yourself with two identical channels by in one a local insert connecting via an XLR cable and in the other adjusting the delay until both signals are in phase again.

Did you write that correctly? 0.31ms is not a relatively high value IMHO.

0.31ms is very low. Conversion takes time. And there are two conversions involved.

Yes, 0,31 ms seemed relatively high to me, because the pure AD-DA conversion is so almost the half of the total latency - despite all the complex calculations such compressions, even DEEP-emulations and the high quality summations.

I would have expected a maximum of 0.2 ms before, but after my simple test, I was wiser.

How would i do that?
Say i send a Drum Channel to the DrumBus (with an external compressor) but i also send it to the Master - so i create some sort of Dry/Wet for that ext Compressor.

I would have to send it to another bus, wich i then delay?

it is quite low, and it’s not a problem unless i would send the unprocessed track to another bus or the master.

If you also need wet and dry, you would have to sacrifice a second bus for dry and delay it accordingly.
For individual channels that are somewhere routed via an external FX and parallel sent to the main, you would probably have to sacrifice a second channel just for the delay to the main.
You could link both channels with ganging and not activate the delay there.

1 Like

I’m late to the party, but I’d say unless you’re doing parallel processing (a drum bus with an external insert plus the drum channels on main bus, e.g.), I simply wouldn’t care.

I often add a small amount of delay to drums anyways. Actually to the whole backline, when I’m mixing in small venues, where direct stage sound and PA sound blend together. This way, the perceived acoustic source of the backline shifts from the PA to the actual backline position on stage (see: Haas effect) and also it sounds cleaner.

1 Like

thats exactly what i want to do, as i’m used to do it in the analog domain. I understand i could send everything intended for parallel processing to a bus wich i would delay accordingly.