MOTUnation.com

Hi -- I'm wondering if anyone can help me figure out a latency problem.

I am using Travelers between a Mac running DP (serving as tape), and an analog Mackie board.

All my monitoring, tracking, and playback goes through the Mackie, just like old times. I am using DP and the Travelers as a substitute for tape, tracking and splicing with them only -- no digital processing.

I am currently submixing 7 recorded drum tracks via one of the Mackie's stereo buses, then routing that stereo submix back into the Traveler, and re-tracking the drums onto 2 new tracks (to save space).

Here's the problem: I noticed all my drum tracks sounding ever so slightly behind, once they'd been bounced down to two tracks.

Using the DP sequence editor, I enlarged comparable tracks and found about 1/1000th of a second delay, between the original and the bounced ones.

Any advice as to how to eliminate this latency? And/or explanation for what is happening?

Is 1/1000th of a sec the time it takes to go out from the Traveler, into the Mackie, and back to the Traveler? Is it something I could avoid by routing my submix differently (or digitally)? Has this always happened in the analog world, and I didn't notice because I couldn't compare two sound waves to the thousandth of a second?

Am I insane to think it is a problem?

I should have mentioned I'm the drummer on this project.

-- Damon

The easiest solution would be to zoom in, and drag the submix files back a few samples so they match the original files.

You could also go under Setup/ConfigureAudioSystem/Fine-TuneAudioI/OTiming and check your latency that way.

Off Topic: What kind of Mackie and outboard gear are you using?

Thank you for the reply -- since I work only in real time I hadn't considered that I/O audio timing idea -- any advice for how to determine the number of samples to move, as I'm working in real time?

Mackie 24x8 board -- outboard gear only various compression, reverb.

If you want to use Fine-Tune I/O (at this point, I wouldn't recomend it. The reason is, it may throw off any overdubs you do in the future. MOTU sets up DP to work perfectly [latency wise] with MOTU interfaces) and you know for sure that it is 1/1000 of a second late, take your sample rate and multiply it by 0.001. That will give you the number of samples to input into fine-tune. Put a "-" sign in front to make sure you're going the right way.

You will probably get better results if you simply grab and slide the overdub files to match the original takes.

Thank you -- sounds like you feel w/real time, I should stick to grabbing and moving the files physically. Seems the easiest way, too. (I think I'll start putting a click down in the front of songs to match up tracks.)

Do you think a latency will occur anytime I re-record something already down in DP? I.e., out from DP, into the board, back to DP.

Or, do you think it's something that happens in the analog world, from routing through submixes like I did with these drum tracks? And I just never noticed before?

exact wrote:Thank you -- sounds like you feel w/real time, I should stick to grabbing and moving the files physically. Seems the easiest way, too. (I think I'll start putting a click down in the front of songs to match up tracks.)

Do you think a latency will occur anytime I re-record something already down in DP? I.e., out from DP, into the board, back to DP.

Or, do you think it's something that happens in the analog world, from routing through submixes like I did with these drum tracks? And I just never noticed before?

It's hard to say if it's your equipment or a computer/interface issue. If I were having a big problem with latency, I would start doing some tests. I'd plug a 1/4" cable into an analog input and output on my interface. Play and record onto another track a sharp transient like a stick click and then zoom in and look at the results. Then, I would take the same signal through just the mixer and back. Record that and check if it throws it off any further. Then I'd add a compressor or eq or both in the channel insert and see if those are pushing the signal back any further. You get the idea. I'm not an analogue gear guru, so I don't know if analog circuits introduce latency. It would make sense if they did. ????

If you find a consistent timing issue, then it's time to go into adjust "Fine-Tune audio I/O".

This reminds me of the days when I would record songs into Audacity in the family PC. I measured the latency of a sound played back through the mixer and headphones and subtracted that much time from each track I recorded. What a pain. But it worked!

This delay is the physical delay introduced by the round trip through the D/A and the A/D. This always happened with Digital tape machines anytime one wanted to do a bounce through the analogue board.

Normally this would only show up if you heard the bounced track and original together as comb filtering.

Congratulations if you can actually notice a 1 millisecond delay, that's a really sharp sense of timing!! (it takes more like 3 or 4 for me to feel something isn't right.)

kind regards

Stephen

Oh and working totally with analogue kit never gave this delay.

Thank you both for your help -- some further tests, and some advice from MOTU, leads me to conclude that my Travelers' A/D and D/A converters are creating a 55 sample delay . . .

Any idea how to calculate that in miliseconds? I'm recording at 44.1k.

Actually the delay is closer to 56 samples, but that would put the overdubs ever so slightly ahead of the track rather than slightly behind, which seems like a worse problem somehow!

I'm fine-tuning the I/O recording offset to compensate, for future overdubs.

It occurs to me that I noticed this only once it was doubled -- that is, only when I bounced a submix back to DP, which effectively moved by another 55 samples tracks that had already (imperceptibly?) been moved by 55 samples . . .

I have to say the whole experience makes me long for analog tape. Will we never all be in that mystical "synch," again?

Digital gives us plenty of reasons to long for tape.

...but hang in there! You'll reap the benefits of digital soon enough.

1000 milliseconds = 1 second. The Math experts may have to correct me on this, but you could use a simple comparison equation.

55/44100(samplerate)=x/1000(miliseconds)

Calculate 55x1000 which equals 55,000. Then divide 55000 by 44100 and get your 1.247 miliseconds or .001247 seconds.

....if its 56 samples, the delay is 1.27 milliseconds. That is not a lot of time.

Like Stephentaylor said, how did you hear that besides playing it back against the original and getting phasy sounds?

Calculate 55x1000 which equals 55,000. Then divide 55000 by 44100 and get your 1.247 miliseconds or .001247 seconds.

....if its 56 samples, the delay is 1.27 milliseconds. That is not a lot of time.

Like Stephentaylor said, how did you hear that besides playing it back against the original and getting phasy sounds?

I didn't "hear" it exactly, I only felt it -- something suddenly felt wrong in the groove of the drum tracks. I'm the drummer on the project, and I had just finished tracking and editing these parts in great detail, listening back to tracks that must be 55 samples off, without ever realizing. But then I bounced down submixes, and without checking those very carefully threw the whole thing into a rough mix. Listening to those roughs (on an iPod, actually), I felt like my drums had gone hazy somehow -- the groove felt just behind. That's when I started trying to figure out if something might have changed in the submix/bounce down.

But at that point, the difference to the original basic tracks would be closer to 3 milliseconds (2 x 55 samples = 2.5 milliseconds) -- which is maybe enough to hear, especially after you've been obsessively thinking about certain snare or ride cymbal hits?

I see. You have a good sense of groove.

I hope all works out well for you!