SOUND ON SOUND

[David_Youngs]

Hello folks,

Could anyone advise on the best way of working out any sample offset between two tracks' audio events inside Cubase please? (or with any freebie plug-in etc.)

My example use case - I want to test the phase coherence (time alignment) of a neat kludge I'm planning to add extra mic pre input capability to my aging Alesis Multimix Firewire 16 interface which is hooked up as my only audio interface to my MacBook Pro. This only has 8 mic inputs with phantom, but it does have another 8 *line* inputs - that I hardly use.

So I currently have on test / sale or return a Behringer Ultragain ADA 8200 ADAT Expander & Mic Preamp unit. The way I want to use the ADA8200 is kind of a neat kludge - by default this unit doesnt pass the signal directly from its Mic inputs to its balanced XLR outputs (see link below)

https://www.sweetwater.com/sweetcare/ar ... g-outputs/

But you can get this functionality using the ADAT output looped back via a short TOSlink cable into the ADAT input on the device - I've tested it and I get audio through the unit in this manner just fine.

I've been reading an interesting thread on the theoretical/actual ADAT processing delay (A>D and D>A) here:

https://www.gearslutz.com/board/music-c ... y-not.html

and it seems logical that there'll be a *very small* processing delay due to the A>D and then D>A conversion in the unit before the balanced output feeds the audio out to my Alesis interface. But given the speed of light and the high speed of conversion I'd *hope* this would be almost non-existent, esp. when compared to the latency of the interface itself.

I wanted to see if I could test and quantify this at all, so had the idea of splitting a signal recorded from one source - say a snare drum hit - by routing it to 2 separate channels, the first being a direct mic input into my Alesis audio interface and onto Cubase , the second being via the ADA8200 mic pre > ADAT loop > ADA8200 output > Alesis line input and onto Cubase as usual.

It then struck me that aside from *looking* at the two waveforms to see (when zoomed in) if there's any visible timing difference/delay, I dont know how to actually *measure* that difference. Is there a way I can use something like a range selection tool or something to see if any delay exists and if it does, how many samples it is?

Any ideas gratefully received, many thanks and Happy Xmas.

David

[Kwackman]

In Cubase you can change the "ruler" from bars and beats, to seconds and fractions of seconds. So, you could probably work it out that way with simple subtraction?

There's probably other ways, e.g. create an event that starts at the beginning of one of the waveforms and ends at the beginning of the delayed version, then look at that event's duration.

But the simple subtraction seems the simplest.

[David_Youngs]

Mmmm, good call on the simple subtraction of *time* with the ruler set to H/M/S instead of bars - I'm so used to working in bars that I've almost forgotten that *real* time exists..

And that's a possible neat idea trimming the event to sample start manually which would work - thanks!

Cheers for the quick reply!
David

[The Elf]

In Cubase you can change the "ruler" from bars and beats, to seconds and fractions of seconds.

In Cubase you can switch the display to samples, so it makes the maths a simple subtraction.

[Hugh Robjohns]

The A>D>A delay will probably be around 1 to 1.5ms at 44.1kHz, and less at higher sample rates.

As long as you're recording different sources through the ADA8200 you probably won't notice in the final mix, although the converter latency might be an issue for foldback monitoring if relevant.

The easiest way to measure the latency is by recording a crisp transient -- a drum hit should be fine -- and the zoom in and use the DAW timeline to measure the distance between the two peaks. Or you could dial in a delay on the earlier track u til the two are aligned.

[David_Youngs]

Thanks Hugh - yup, you'e absolutely right. I tested this using a split signal as descibed above and the A>D>A loopback/passthrough delay is sub 1mS. Sorry, cant seem to figure out how to post the screenshot I took, but the range selection definitely shows under 1mS offset between the 2 tracks.

Cheers for your help everyone!
David

[Hugh Robjohns]

:thumbup:

1 millisecond is about the same delay as you get if you pulled the microphone 12 inches further back from the source... so not really a problem in the context of a mix in most cases.

That said, a 1ms delay in the foldback to headphones is on the edge of upsetting some people if they're singing/talking as it causes comb-filtering with the direct sound they hear .