| I got a question about latency compensation in PTLE via PM from this thread and I thought I'd share my response here for the general public good...
When I do an outboard pass on an individual part I bring it back in printed on a separate audio track and then I move it back in time to align with the original signal by eye, zooming in on the waveform. Then I inactivate the old track and mix from there with the new one.
When I'm trying to do real-time mixing with hardware inserts (or latency-including plugins) I load up ping regions (that I hand-drew with the pencil tool) into every audio track out past the end of the song, aligned using grid mode. Then I record those pings back in to my mixdown track (I record to disk rather than bounce to disk) and I see whether they are aligned or not (I set reverbs/delays to 100% dry for this). I time delay using the time adjuster plugin everything that comes earlier than the slowest of the pings, sometimes on the individual tracks, sometimes on busses. I sometimes need a couple layers of delays in order to align sends/returns when I have latency on the busses.
This is one reason why I generally just print back hardware passes and manually realign (the other reason is recall). And when using UAD or Duende I will try to make sure there's an equal number of those plugs everywhere so that they all have the same latencies.
And I'll reiterate that most DAWs have this problem even if they claim automatic latency compensation, and this is the general solution. This could be automated and a few DAWs do automate via pings. I predict all will by the end of the decade but you're going to have to really bear down on the programmers for this it appears. No idea why this wasn't solved generally 10 years ago, but probably most of the DAW coders are doing electronica or something where phase is less of a problem. |