It’s fine to use limiters on individual sources and subgroups — but in this context, you’d use them very differently from typical mastering limiters.
Is it ever appropriate to place a limiter on a routing folder for specific parts prior to them hitting the final mix bus? The idea here is to get the parts (drums, guitar, keys, and vocals) up to a loudness level I’d like, then push them the final few decibels in the mix bus before hitting the mastering chain. Sonically, the main problem I’ve encountered is a consistent need to re‑transient‑shape the drums, as they lose punch and energy as the limiter squeezes them into place. What other issues might this method cause?
Anon via email
Mix Expert Mike Senior replies: Reading between the lines, I think you might be conflating two different types of limiting. The first type is a mixing limiter, which is effectively just an infinity‑ratio compressor designed to deal with balance issues with maximum speed/firmness. I regularly use this kind of limiting on channels and/or submixes whenever I want to control peak levels but, crucially, my aim in those circumstances is to achieve a specific change to the musical balance of whatever signal I’m limiting. For example, I’ll often use a few decibels of limiting with a medium release to even out inconsistent live snare and kick tracks without appreciably affecting their attack/sustain envelope. Or I might apply limiting with a faster release to reduce the onset spike of hand claps, percussion, finger clicks, cowbell or woodblock, to make more of those instruments’ sustain in the mix. Limiting instruments and submixes in this way is just an extension of compression, so I have no problem using limiting if I feel the musical balance demands it — it’s something that’s much more likely on individual channels than on submixes in practice.
What I don’t do at mixdown, though, is use the other type of limiter — namely a mastering limiter — to try to achieve some kind of target dB LUFS reading or peak‑to‑average ratio for individual tracks or submixes. Mastering limiters are usually much more sophisticated under the hood than mixing limiters, often with complicated multiband facilities and saturation/clipping functions, and their purpose isn’t really a musical one. They’re more concerned with the technical goal of increasing subjective loudness within a fixed digital headroom, and there are several reasons I wouldn’t recommend using them at mixdown.
You’ve already highlighted one reason: this kind of limiting removes too many of a signal’s level peaks too early in the signal chain, which makes percussive instruments feel like they’ve lost edge. Now, you might argue that “the limiter at the mastering stage will do that too, so what’s the difference?” Well, the difference is that once those peaks are gone, they won’t trigger your main master‑bus compressor in the way they normally would, meaning you’ll lose some of its nice gluing/pumping side‑effects. (Though, to be fair, this problem may also occur if you use mixing limiters too aggressively as well.)
But you also need to realise that mastering‑style limiters that aim to boost loudness within a fixed digital headroom often end up flattening the tops of the waveform shapes with some kind of clipping, and they may also use complicated phase‑rotation techniques to reduce peak levels by shifting the phase of mix harmonics relative to each other. Both these underlying processes have (usually undesirable) transient softening/smearing side‑effects, which is the price you pay for trying to increase subjective loudness without increasing peak signal level — a trade‑off most people are happy to make to some extent at the final stage of the mastering process.
It’s better to avoid using mastering limiting at mixdown.
However, if you process, say, your drums submix this way during mixdown, then any further phase‑shift incurred by subsequent downstream mix processing (perhaps on the master bus) has the potential to reintroduce the signal‑level peaks you were previously trying to control. (Bear in mind that most EQs cause phase‑shift, as do plenty of analogue‑modelled plug‑ins.) If that happens, you’ve then lost some of the benefits of the limiting (ie. the control of peak levels), but won’t have lost any of the unwanted side‑effects, so you’ll incur additional processing side‑effects when those are again limited at the mastering stage. So it’s better to avoid using mastering limiting at mixdown, just so your mix only has to suffer its damaging side‑effects once over.