Compensate latency when recording samples?

Help, Support & Bugs Beginners Questions - Renoise
1

Simple question: when should I enable the “compensate input and output latencies” option, when recording a new sample while playing my tune?

My sound card enables “direct monitoring”… where I can hear what I’m playing, without going through renoise, and then mix the renoise output with it. I think that’s the case where I want to enable i/o compensation… right?

But if I don’t use the direct monitoring, and instead listen to the input through renoise when recording (the little speaker icon) then I think I wouldn’t want to use i/o compensation for the recording.

Am I right or wrong about that?

2

bump :slight_smile:

maybe my timing just sucks anyway and so it doesn’t really matter… but I’d like to know from an ideal standpoint whether I should “compensate input and output latencies” when monitoring via Renoise, or not.

3

okay, I devised a little test to help me figure this out on my own. I increased the device settings latency preference to its maximum of 93ms (that’s the max with my config, anyway). Then I recorded a series of quarter-note hi-hats played along to the metronome, with monitoring disabled so the sound wouldn’t throw off my timing.

If you monitor through Renoise, you want to enable “compensate input and output latencies” in the sample recording dialog.

I now believe this conclusion to be false (leaving it in for posterity): If you monitor direct through your sound card, you might want to disable that option… but I didn’t test that.

I also learned that my timing is bad enough that when I use my usual latency setting of 8ms, it doesn’t really matter :slight_smile:

4

Hmmm, i assumed that the latency compensation simply calculated or in some way found out the actual latency from input to recording and then simply cut that off at the beginning of the recording?

5

Hmmm, i assumed that the latency compensation simply calculated or in some way found out the actual latency from input to recording and then simply cut that off at the beginning of the recording?

Basically, yeah. It does that based on the audio device and/or the latency preference… I’m assuming :slight_smile:

I know what it does, I just didn’t know when I want to use it - thus me asking, and then devising a test for it.

6

You could calculate it and remove it manually, but i don’t see any advantage of doing it so, i would just say keep it ticked on either way you monitor the input.

If you use direct monitoring, then, assuming you record dry, you will get just about the exact result as what you monitored while recording. In theory at least if i’m not missing something.

7

I think you misunderstand me… I asked because I didn’t know. Nobody responded. I came up with a way to figure it out myself, and I did :slight_smile:

There’s no more theory or speculation for me. I ran an experiment, and figured out the answer to my question.

8

yes, but you said: If you monitor direct through your sound card, you might want to disable that option… but I didn’t test that.

that’s what i responded to. :wink:

I just tried to see why you would want to disable it when direct monitoring as it doesn’t make much sense.

9

yes, but you said: If you monitor direct through your sound card, you might want to disable that option… but I didn’t test that.

that’s what i responded to. :wink:

I just tried to see why you would want to disable it when direct monitoring as it doesn’t make much sense.

Here’s why I think there’s a difference, and why I posted the original question (and soon you’ll see why I think I came to the wrong conclusion…)

When monitoring through Renoise, there’s some latency - whatever’s set in the preferences. When monitoring direct through the sound card, there’s no latency. So it makes sense to me that in one setup, you’d enable latency compensation (in the “record new sample dialog” that is!!), and in one setup, you’d disable it.

The question is… which option for which setup?

In my original post, I said: “if I don’t use the direct monitoring, and instead listen to the input through renoise when recording (the little speaker icon) then I think I wouldn’t want to use i/o compensation for the recording”

so, original hypothesis: direct monitoring, enable latency compensation; software monitoring, disable it

Because when direct monitoring, the sound I hear direct through my sound card will avoid the latency, and thus Renoise will need to compensate for it…

so then I made my experiment. Here’s where I went wrong.

By disabling monitoring altogether and playing silently to the metronome, I introduced a condition that never happens in practice. When monitoring through software, you naturally play a little bit ahead to make up for the latency. This was the reasoning behind my original hypothesis… and is hard for me to test. If I disable monitoring, then I can play along to the metronome - and see that my uncompensated hits are behind the beat. But in practice, I will have monitoring on… and will play a bit ahead so that what I hear matches the beat.

So I’m back to square one here. I think that with software monitoring enabled, I want to disable latency compensation, because I compensate for the latency naturally through my own playing. And when direct monitoring, I want to enable latency compensation, because what I hear coming through the sound card will be slightly out of sync from Renoise’s perspective.

In practice, with a latency of 8ms, it doesn’t seem to matter, because my body has more than 8ms of jitter :smiley:

I’ve just done a bit more testing, with software monitoring enabled, at a slower tempo… with latency compensation disabled, the sample is consistently behind the beat (with space before the first beat). With latency compensation enabled, the sample is consistently ahead of the beat (with the first beat partially chopped off. This was with a device latency preference of 24ms.

When I set the device latency to 8ms, I can’t really tell a difference… part of the first beat is still chopped off. Uncompensated seems to have a bit better timing, but not by much.

So in summary… I have no idea wtf is going on :smiley: It seems like, for my combination of playing ability + sound card + renoise settings, the answer is to slice and sequence if timing matters, rather than relying on my hand-played sample. If I don’t want my sample getting chopped off at the beginning, then I need to play two loops around, so I can hand-trim it at the right place.

Two things I’ll be messing with: 1 - slice in renoise, and sequence it right on the beat. 2- record in to ableton, and use its warping / quantization to get the sample where I want it, and then bring it in to Renoise. Neither is ideal to me, but it’s pretty clear now that I can’t rely on Renoise’s pattern-sychronized sample recording.