Old 10-13-2019, 10:05 AM   #1
JohnnyMusic
Human being with feelings
 
JohnnyMusic's Avatar
 
Join Date: Sep 2014
Location: Twin Cities, Mn
Posts: 233
Default How to calibrate latency manual offset

I've been using the driver reported latency and not bothering further with it but decided to fine tune my latency, since I am doing a lot of tracking for a project. My primary question is how to determine what to set the manual latency offset at based on my test and whether to use input or output offset.

I've tried searching for this but couldn't find a clear answer to my questions. I also asked the question on a thread here but got no response. I am hoping someone that knows the answers can help clarify.

I tested by routing the inserted click source track out of my interface (a motu 828x) and back into three different inputs to record it: The 828x built in pre-amp, 828x without the preamp (analog in), and motu 8 pre expansion converter with built in preamp. The latencies I got were pretty small: 153, 133, and 150 samples respectively. Is the lower latency with the no preamp test because the preamp adds latency?

In the case of audio, (I monitor through my interface for recording guitar, bass, vocals etc), so my question is:
1) should I take 1/2 of the 133 samples (66.5) to use for manual offset, to compensate for the playback latency since when I am recording, since I am only listening to playback of the recorded tracks through reaper and not input or does the time for input also need to be included because it is taking this extra little bit of time to process the audio input (bass for example- confusing!?

2)As an aside I noticed recently that when I recorded bass di and also through a sansamp , the sansamp signal was recorded slightly later (can't remember how many samples right now but I can check)Edit: it was about 70 samples later. Edit 2: I should note I had the signal go through a tuner first then out the tuner out to the sansamp and the parallel out on the tuner to the interface, so I shouldn't assume those 2 outputs are the same.

3) Should I use input or output offset (I tend to think output since it is just with playback but I don't know if that makes sense)?

4) I haven't tested midi yet but I will be tracking some parts by playing through a VST for some tracks, any advice for latency in that situation? Will I want to use both the input and output latency since I will be monitoring the VST live?

Thanks for any advice!
John
JohnnyMusic is offline   Reply With Quote
Old 10-13-2019, 10:50 AM   #2
Pickasso
Human being with feelings
 
Join Date: Aug 2009
Posts: 204
Default

*Insert/click source
*Record playback source thru input source. (when recording drums, I'll put headphones next to mic, when recording guitar I'll put headphones directly on top of pickup, crank it loud)
*Uncheck snap enabled
*Normalize file, or raise the file volume under item properties so you can see the peaks.
*Measure start of click beat to start of recorded click in samples using mouse as time selector.
*Enter the amount of samples measured here: options/preferences/recording/output manual offset (in the samples box)

Repeat, keep adjusting until happy. Every interface is different, usually you can get it really close. Most cases keep the 'use audio driver reported latency' active. If nothing else works, last resort, try without that.

Takes a while, but once you get it really close, and repeatable, tracking becomes so much tighter and more professional sounding!
Pickasso is online now   Reply With Quote
Old 10-13-2019, 11:23 AM   #3
JohnnyMusic
Human being with feelings
 
JohnnyMusic's Avatar
 
Join Date: Sep 2014
Location: Twin Cities, Mn
Posts: 233
Default

Hey Pickasso, thanks for replying.
That is pretty much what I did and got the samples of offset as I noted.
So you are saying to use the entire round trip latency in the manual offset.

I am a bit confused but I though maybe I shouldn't use the whole round trip offset.

Since I am only monitoring prerecorded material and not my bass. Also should the offset be input or output?

Thanks!
John
JohnnyMusic is offline   Reply With Quote
Old 10-13-2019, 11:42 AM   #4
SonicAxiom
Human being with feelings
 
SonicAxiom's Avatar
 
Join Date: Dec 2012
Location: Germany
Posts: 1,946
Default

Quote:
Originally Posted by JohnnyMusic View Post
I've been using the driver reported latency and not bothering further with it but decided to fine tune my latency, since I am doing a lot of tracking for a project. My primary question is how to determine what to set the manual latency offset at based on my test and whether to use input or output offset.

I've tried searching for this but couldn't find a clear answer to my questions. I also asked the question on a thread here but got no response. I am hoping someone that knows the answers can help clarify.

I tested by routing the inserted click source track out of my interface (a motu 828x) and back into three different inputs to record it: The 828x built in pre-amp, 828x without the preamp (analog in), and motu 8 pre expansion converter with built in preamp. The latencies I got were pretty small: 153, 133, and 150 samples respectively. Is the lower latency with the no preamp test because the preamp adds latency?

In the case of audio, (I monitor through my interface for recording guitar, bass, vocals etc), so my question is:
1) should I take 1/2 of the 133 samples (66.5) to use for manual offset, to compensate for the playback latency since when I am recording, since I am only listening to playback of the recorded tracks through reaper and not input or does the time for input also need to be included because it is taking this extra little bit of time to process the audio input (bass for example- confusing!?

2)As an aside I noticed recently that when I recorded bass di and also through a sansamp , the sansamp signal was recorded slightly later (can't remember how many samples right now but I can check)Edit: it was about 70 samples later. Edit 2: I should note I had the signal go through a tuner first then out the tuner out to the sansamp and the parallel out on the tuner to the interface, so I shouldn't assume those 2 outputs are the same.

3) Should I use input or output offset (I tend to think output since it is just with playback but I don't know if that makes sense)?

4) I haven't tested midi yet but I will be tracking some parts by playing through a VST for some tracks, any advice for latency in that situation? Will I want to use both the input and output latency since I will be monitoring the VST live?

Thanks for any advice!
John
it seems that you did the loop back test correctly: running the glued click source item's signal out of your interface and directly back into one of its inputs and then re-record this signal. After that, measure the offset between the waveforms (original, glued click source vs. re-recorded input signal). Disable "use driver reported latency" and put the found sample offset value into the manual latency offset (input, samples) field. I'm not sure wether putting it into the "output" field will make any difference. I'm using the input field.

As a result, your future recordings through the measured interface will be recorded at the correct spots on the timeline in Reaper. I can't say why latency varies depending on the type of input you choose. Probably the MOTU employes some audio buffering in some of its signal paths. I'd use the latency value of the input which I will be using most of the time. The different latencies will probably never be a problem unless you are recording one source (like drums) spread across more than one input type simultaneously. In this case, there will be phase delays introduced between tracks that should be remedied by using the "JS: time adjustment" plugin.

I have found no latency difference when using any of my analog preamps vs. the converter's input so (entirely analog) preamps seem to not introduce additional latency.

.
__________________
[Check out my free VST plugin collection here.]
SonicAxiom is offline   Reply With Quote
Old 10-13-2019, 12:52 PM   #5
Pickasso
Human being with feelings
 
Join Date: Aug 2009
Posts: 204
Default

Quote:
Originally Posted by JohnnyMusic View Post
Hey Pickasso, thanks for replying.
That is pretty much what I did and got the samples of offset as I noted.
So you are saying to use the entire round trip latency in the manual offset.

I am a bit confused but I though maybe I shouldn't use the whole round trip offset.

Since I am only monitoring prerecorded material and not my bass. Also should the offset be input or output?

Thanks!
John
*You're trying to measure what comes out of the headphones to your farthest input. I.E. mic or guitar pickup, that way you miss nothing in the chain.
*Should be 'output manual offset'.
*I don't disable the reported latency as SonicAxiom suggests. It's more consistent to leave it active. You're just trying to adjust how much it's off.

Try any or all of the above, most interfaces you'll get there.
Good Luck.
Pickasso is online now   Reply With Quote
Old 10-13-2019, 06:17 PM   #6
JohnnyMusic
Human being with feelings
 
JohnnyMusic's Avatar
 
Join Date: Sep 2014
Location: Twin Cities, Mn
Posts: 233
Default

Hmmm...
-I did my tests with the "use driver reported latency" ticked or on.

-I thought what I was doing here was measuring the slight inaccuracy in that driver reported latency and then compensating for it with the manual offset.

-I entered the reported latency in offset and it did get it to 0 samples of offset.

-My question also though was do I still need the whole round trip number if I am monitoring my bass through the hardware or maybe half of that? I guess once I play a note on the guitar or whatever, There is some latency as it travels back to the reaper and gets recorded, as well as the recorded material that I am playing to, taking time to get sent from reaper, and out to the speakers. Maybe I'll just try the shortest round trip latency that I found, experiment with that, and go from there.

Thanks again for the explanations.
John

Last edited by JohnnyMusic; 10-13-2019 at 06:28 PM.
JohnnyMusic is offline   Reply With Quote
Old 10-16-2019, 08:24 AM   #7
Goldreap
Human being with feelings
 
Join Date: Feb 2009
Location: UK
Posts: 1,054
Default

For the purpose of recording audio while listening to a click REAPER just needs to know the actual total IN + OUT latency. Whether you enter input or output offset doesn't matter.

But input vs offset can become relevant when dealing with MIDI. Consider the following:

In order to grid sync both MIDI and audio input, REAPER treats MIDI input latency as being equal to audio input latency.
An audio device might report correct total latency but mess up the relative amounts of input/ output.
Bear in mind that output offset moves both MIDI and audio in tandem but input offset only moves audio. This allows you to juggle them to sync both MIDI and audio to the grid.
E.G. if MIDI is landing early (but audio is OK) it would seem that the audio device is over-reporting input latency. A fix might be to try a negative output offset (both MIDI and audio will land later on the grid), but also enter a positive input offset (same amount) to keep the audio where it is.
Goldreap is offline   Reply With Quote
Old 10-16-2019, 05:48 PM   #8
JohnnyMusic
Human being with feelings
 
JohnnyMusic's Avatar
 
Join Date: Sep 2014
Location: Twin Cities, Mn
Posts: 233
Default

Quote:
Originally Posted by Goldreap View Post
For the purpose of recording audio while listening to a click REAPER just needs to know the actual total IN + OUT latency. Whether you enter input or output offset doesn't matter.

But input vs offset can become relevant when dealing with MIDI. Consider the following:

In order to grid sync both MIDI and audio input, REAPER treats MIDI input latency as being equal to audio input latency.
An audio device might report correct total latency but mess up the relative amounts of input/ output.
Bear in mind that output offset moves both MIDI and audio in tandem but input offset only moves audio. This allows you to juggle them to sync both MIDI and audio to the grid.
E.G. if MIDI is landing early (but audio is OK) it would seem that the audio device is over-reporting input latency. A fix might be to try a negative output offset (both MIDI and audio will land later on the grid), but also enter a positive input offset (same amount) to keep the audio where it is.
Helpful insights! Thanks!
JohnnyMusic is offline   Reply With Quote
Old 10-17-2019, 07:25 AM   #9
davetbass
Human being with feelings
 
Join Date: May 2015
Posts: 155
Default

I found this video to be fantastic and it shows a quick way using built in plugin:



https://www.youtube.com/watch?v=TzS--D765Zw
davetbass is offline   Reply With Quote
Old 10-17-2019, 06:44 PM   #10
JohnnyMusic
Human being with feelings
 
JohnnyMusic's Avatar
 
Join Date: Sep 2014
Location: Twin Cities, Mn
Posts: 233
Default

Quote:
Originally Posted by davetbass View Post
I found this video to be fantastic and it shows a quick way using built in plugin:



https://www.youtube.com/watch?v=TzS--D765Zw
I tried the method in this video but I couldn't get it to work, not sure why.
Recording the click and zooming in and use time selection to count the samples for me was easy enough though.
JohnnyMusic is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 10:39 PM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, vBulletin Solutions Inc.