Another try at the latency test

I’ve been using Time Shift to line up my overdubbed tracks but it’s very imprecise, so I decided that once and for all I’m going to perform a proper latency test. I understand the procedure and I can generate a rhythm track perfectly, but when I go to record the next track it doesn’t do anything. I’m supposed to see two tracks of the rhythm track with the second one lagging behind, but all I see is a straight line.

What am I doing wrong?
Screen Shot 2020-09-25 at 4.03.33 PM.png

The top track would be your “backing track” or guide track. You’re supposed to listen to that while you sing or perform. Anything that requires a microphone also requires that you use headphones.

The simplest process is jam your headphone against your microphone.

No, I don’t recommend using Apple earbuds to shoot sound. They were handy for the picture.

You should get a new track with the ticks slightly distorted from the round trip, but exactly one latency off.

Count the time and combine that time with the latency time in Audacity > Preferences > Devices > Latency.

Check it again.

You are required by state law to go the wrong way the first time and make it worse. You should know that.


Did that and it worked, maybe too well. After recording with my mic jammed into my phones, the top and bottom tracks lined up perfectly. What the hell…anyway can I proceed from here?

If so then how do I proceed? Can I start recording and overdubbing on new tracks and everything will be in sync? What do I do with the latency test, record over it, leave it alone, delete it? Do I run a latency test each time I’m about to record?
Screen Shot 2020-09-25 at 5.22.50 PM.png

After playing with latency all afternoon I’ve come up with a routine but I’m not sure if it will work this way.

  1. I generate a rhythm track.
  2. I put my mic in my phones and record another track under the rhythm track.
  3. I amplify the second track and use the time shift to pull the bottom track into alignment with the rhythm track. It’s as close to perfect as possible.

Will this method correct the latency?

The object of the latency correction is to never use the Time Shift Tool again. If you perform in time with the backing track (and everything else goes OK), your performance will come out in perfect sync, automatically, as many different instruments as you wish (up to the power of the computer).

With reference to the top illustration. You got lucky/unlucky. Your performance is exactly one click off from the backing track.

Screen Shot 2020-09-25 at 5.15.15 PM.png
The red bracket points to the two clicks that are supposed to line up. The circle on the left? Those are not the same click.

So figure out the difference between those two and combine it with the existing latency setting. Once you throw off your accidental match, you’ll be able to pick out a timing difference much easier.

This sound clip is a more normal close but no cigar.


So I’m right back where I started. I can generate a rhythm track perfectly, but when I go to record the next track it doesn’t do anything. I’m supposed to see two tracks of the rhythm track with the second one lagging behind, but all I see is a straight line. When I amplify that line all
I get is noise.

Hell, let me try it again. I’m pessimistic but I’ll try the thing again.
Screen Shot 2020-09-25 at 9.12.51 PM.png

Ok! Where the hell does the backing track come from? All I have at this point is an incomplete latency test, with the second track containing nothing but noise. I’m stuck, I’m completely stuck

Is this your mixer?
Screen Shot 2020-09-25 at 6.20.51 PM.png
Before you get your forehead even more bloody: Basic troubleshooting tells us to follow the signals down the pathway rather than plucking it out of the middle with no context.

Do you get the SSL bouncing sound meter when you speak? Can your hear yourself at the SSL headphone connection when you perform?

Without digging in the instructions, I’m guessing the mix knob, upper right, is supposed to fade between the existing track (backing?) and your performance. Does it?


The SSL2 is my interface. It’s the one below the SSL2+. My mixer is the Allen and Heath ZED6FX.

Headphones are plugged into the interface. Yes, I can hear myself when I sing or play.

Yes, that’s exactly what the monitor mixer does.

Would it help to see the stuff I’m using?
my stuff.png

Is that 77 keys? I had one of those before my current 88. I have a fantasy of firing them both up at once. I know someone who did that. He had a stacking arrangement like a large pipe organ.

You have one of those sit-in-the-corner-with-strong-coffee problems. If you can’t record your SSL, how did you get your posted tests where the latency is one click off?


Sorry, I don’t understand what you mean by not being able to record the SSL. The mixer goes into the SSL which goes into the computer’s UBS input.

I used the mic jammed into the phones method you mentioned earlier to get those tests. The mic is going into the mixer, do I want the mic to be plugged directly into the interface? I can do that.

Yup, 77 keys, I think. I’ve only used it for composing, I’m not much of a keyboard player. Do you have an Ensoniq?

Here are pictures of what happens when I run the test.

  1. The two out of alignment waveforms (after placing the mic into the phones).

  2. The space between the pings. I couldn’t make them the size shown in the manual because I couldn’t get two pings on the screen at the same time. Should still work, no?

  3. The result after I “Delete the second track by clicking the close box X in its control panel”, “Click the Fit To Width” button to see the entire click track, "click in the Track Control Panel of the remaining track to select it, then clicking the record button, and “after the recording is finished.”

So something must be wrong somewhere in the third stage, maybe, I don’t know.

If insanity is doing the same thing over and over and expecting a different result then I’m certifiable.