Your browser is out-of-date!

Update your browser to view this website correctly. Update my browser now


Time Is Not On Our Side

What’s the earliest audio device you can think of where signal latency was an issue in its normal use?

What’s the earliest audio device you can think of where signal latency was an issue in its normal use? If you are automatically thinking digital, you are likely younger than I am. If you thought, “multitrack analog tape recorder,” then we are on the same track.

Surely most all of you know that when you are overdubbing on an analog machine, you have to listen to your previously recorded tracks off of the record head so that your new tracks and punch-ins are in the same time alignment as the earlier tracks. Analog recorders’ playback heads being physically spaced a distance past the record head in the tape path, if you monitored off the play head during recording, the new tracks would be separated in time by the tape speed divided by the distance between the heads. That, my friends, is analog, mechanically induced latency.

When working with multitrack digital tape recorders (you know, those washingmachine- sized devices you now see in studio corners and used as shelving for stacking hard drives), you couldn’t blend a signal monitored through a tape machine with a live signal running through the analog consoles without comb filtering occurring. The tape machines automatically took care of lining up new tracks and previously recorded tracks in time within the recorder’s electronics and on tape, but the trip to the recorder and back still took time — whether tape was rolling or not. Tracks monitored through the recorder had at least the conversion latency, ADDA, to contend with.

On tape-based digital multitracks — no oversampling converters, analog filtering — those latencies were fairly tame, certainly well less than the 2 mS that we considered at our studio to be the maximum acceptable ceiling — a “safe” roundtrip latency between a singer’s voice and his/her headphone mix that would annoy only the very few and most discerning vocalists. We considered digital consoles of the day unsuitable for tracking, in part because they added unacceptably to the latency overhead. Even some analog console operations could be problematic as your cue mix had to come from, and through, the tape machine.

That made unusable some console smart cueing features that blended playback and analog path tracks. It also created one nightmare of a session I remember far too well, when a brand new digitally controlled analog console was being used for a tracking date; the monitor path and channel path controls were layered, and the engineer added elements from both direct and tape return paths to the headphone mix, creating a confusing mess for the musicians.

In our mostly digital production flows, we have become accustomed to coping with latencies, along with the intimately related topics of interchannel track alignment and processing latency compensation. Modern recording systems, by necessity, have a variety of schemes allowing zero or low latency monitoring paths for tracking and overdubs. While we can now approach latency issues with an informed confidence, as a routine part of modern recording as well as live sound, be thankful to those who led the way in the digital revolution, as this was not always the case.