Finagle My Song · Blog
Limiter vs clipper for loud masters
Everyone reaches for a limiter. It's the transparent choice, the safe one, the thing that doesn't distort. So when a master comes back loud but lifeless, nobody suspects the limiter did it.
It usually did.
The two tools do genuinely different things
A limiter reduces gain. A peak comes through, it turns the level down, then lets it back up over a release time. That release is the part nobody thinks about. Pull down a peak with a 200ms release and you've turned down everything around that peak for 200ms, not just the peak itself.
Now do that a few hundred times across a track to hit a loud target. You're not shaving transients any more, you're modulating the level of the whole song. That's pumping. It's also why macro-dynamics quietly vanish, because a limiter can't tell the difference between a snare hit and a section that's simply louder than the one before it.
A clipper doesn't touch gain. It truncates the top of the waveform, sample by sample, and leaves everything below the threshold sitting exactly where it was. Your quiet intro stays as quiet, relative to the chorus, as you wrote it.
The cost is distortion. Truncating a waveform generates harmonics, and that's the whole reason clipping carries the reputation it does.
Why that reputation is only half right
Those harmonics on dense, aggressive material read as grit and energy. The same harmonics on an exposed acoustic guitar read as a broken file.
So "which one is better" turns out to be the wrong question. What matters is what your material can absorb, and how loud you actually intend to go.
What it looks like on a real track
A song with a soft electronic intro that opens into a full band at 0:45. Target: -9 LUFS. Same settings both times, only the final stage changed.
With the limiter: the soft intro picked up pumping and over-processing artifacts, and the waveform went flat and full-height from end to end. Intro and band ended up at almost the same level, which killed the one piece of expansion the whole arrangement was built around.
With the clipper, at the same -9 and the same dials: artifacts gone, and the intro-to-band expansion survived.
On dense material the clipper also held onto roughly 3 LU more loudness than the limiter before anything audibly broke. Same target, and it kept more of it with fewer artifacts.
So when do you use which
Clipper when the material is dense and you genuinely want it loud. Aggressive electronic, rock, hip-hop, anything built on energy rather than air. It buys you loudness without paying in macro-dynamics.
Limiter when the material is sparse, clean, or dynamic by design, and when you are not pushing for a loud target in the first place. Acoustic, jazz, ambient, anything where an exposed instrument sits alone in the mix.
The mistake is reaching for the limiter by default because it's the "safe" one. Push for a loud target and it's the expensive choice.
Don't forget true peak
Either way, the file still has to survive lossy playback. Inter-sample peaks distort on streaming even when your sample peaks look perfectly clean, so measure the ceiling with oversampling rather than trusting the peak meter in your DAW.
Sensible ceilings: about -1.0 dBTP on the limiter path, about -0.5 dBTP on the clipper path, both measured with 4x oversampling. Clip without checking true peak and you're shipping distortion you never heard.
Try it on your own track
Here's the practical version. Take something you already mastered loud and look at its DR. If that number is low and the track feels tiring rather than energetic, odds are the limiter made a decision you didn't.
That split is why the mastering tool I built has a Crunch toggle. Off, it runs a true-peak limiter at -1.0 dBTP. On, the final stage becomes a 4x oversampled true-peak clipper at -0.5 dBTP, which is the loud path. Everything else in the chain stays the same, and you can A/B the two at matched loudness so you're judging artifacts instead of volume. Upload a track at Finagle My Song and hearing either version is free, no account.
Whichever tool you land on, pick it on purpose. Loud is a decision. Flat is what happens when you don't make it.