Note that the master fader is set to 0.0, and the output has peaked at –6.2.
I get lots of contradictory advice for setting levels with digital recording—from “as close to 0 as possible” to -18dB,andeverywhere in between. So . . . what is it?
It depends on where in the signal chain you’re checking levels. When entering your interface inputs, leave some headroom to accommodate unanticipated transients; typically, keep peaks below –6dB (although many engineers are more conservative). While giving up 6dB of dynamic range sacrifices about one bit of resolution, with today’s 24-bit converters this makes no practical difference.
Once within your DAW, an audio engine with at least 32-bit floating-point resolution offers virtually unlimited dynamic range. However, although individual channel meters can “go into the red” without causing distortion, dynamic range is again an issue when these signals leave your DAW and return to the audio interface. One potential problem is inter-sample distortion where, due to the output-smoothing process, the true signal level can exceed what the meters indicate (sometimes by several dB). So, when mixing, it’s good practice to keep output peaks lower than –6dB—then during mastering, raise the level as appropriate while checking for inter-sample distortion.
Finally, there can be a tendency toward “channel level creep,” which necessitates turning down the master output level. It’s better to keep the master around 0 and if needed, group the channel faders temporarily, and adjust their levels so the output signal peaks at around –6dB.