Syntrillium Backup Forums :: View topic

clothesburner · Location: USA

I don't know what got me on this latency kick again, but something did, and I'd like your opinions concerning my findings.

Test #1 - most of the time when recording a band, I'll track the drums, bass, and guitar first and then go back and add vocals. Well, I started being concerned that the singer's performance, since it was being tracked apart from the other instruments, was not lining up correctly with the other tracks whether it was due to a little latency on the output side or latency on the input side. To test this, I recorded a metronome track to track one of CEP which represents the performance mixdown to which the singer will perform. Then, on track to, I held a microphone up to my monitors and recorded the metronome track playing through them - the metronome playing through the speakers represented the signal coming through the headphones and the microphone represented the input through which the singer sings. My finding was that the metronome track that I recorded through my monitors did not line up with the metronome track I had recorded on track one - the track I recorded through my monitors was exactly 1 ms off.

Now, I understand completely that it takes time for sound to travel into the microphone so I made certain to hold the mic right on the speaker.

Where is this latency coming from? When I do overdubs, do I need to negatively delay the overdub so that it falls right in sync with the original recorded tracks or am I wasting my time worrying about it? To be quite honest, I've never once heard this 1 ms delay and think I'm being quite anal - but if there's a way to fix this 1 ms delay situation, I'd like to know.

Test #2 - another thing I do a lot of the time is run my direct guitar signals out to my POD and back in. For this experiment, I loaded a direct guitar signal on track one of CEP, ran it through the POD, and re-recorded it on to track two. The setting for the POD was something clean (as opposed to something real effects heavy) so that the signal wouldn't be changed much going back in. And my finding was that the newly recorded signal was exactly 6 ms off of the original. Same situation as above: negatively delay the re-recorded track so that it is in perfect sync with the original track.

Just so you know, I was using a Delta 1010 for this experiment with the buffer setting in the M-audio control panel set to 2 ms, the lowest latency setting offered.

Is the 1 ms offset from test one worth worrying about when doing overdubs? Is there a way to prevent it? I honestly can't live with the 6 ms delay from test two so, from now on, I'll be negatively delaying either the re-recorded signal to line it up with the original or the orignal recorded signal before it ever goes out.

Thank you for your responses to my post.

AMSG · Location: Sweden

I assume this delay can be caused by your own workstation, the software and of course the fact that it takes time for sound to travel. But I'm not an expert on this, I guess others here will be able to explain all of this in detail for us=)

And I myself don't worry about such small delays as 1ms and 6ms. Especially since I often use delay to get a wider and bigger sound. Even 6 ms won't be audible for us, it'll sound in sync anyway.

djwayne · Location: USA

There's a way to get it right on. First record, your click track on the first channel, and then record your first track in time with that as you hear it, next record the following track also in time with the click track, and take one and take two and so on should be inline. Then remove the original click track in the final mix.

Get it too perfect, and it'll sound like a machine, with no human touch or feel to it.

clothesburner · Location: USA

Thanks for the click track idea. That makes perfect and logical sense.