Learn
What is frequency masking?
Frequency masking is the reason a busy arrangement can sound cluttered even when every track sounds great on its own. When two sounds occupy the same part of the spectrum at the same time, your ear stops resolving the quieter one — the detail is still there, but you can't hear it.
The short version
Human hearing has limited resolution. When a loud sound and a softer sound share a frequency region, the louder one raises your hearing threshold in that region — so the softer sound falls below what you can perceive. It hasn't been removed from the signal; your ear simply stops picking it out. In a mix, that's heard as mud: instruments lose definition, and reaching for more volume or more EQ only moves the problem around.
Masking is a property of hearing, not of your speakers or your DAW. That's why it survives a great recording and a clean signal chain — and why you can't fix it by turning things up.
Why it happens: the basics of psychoacoustics
The inner ear analyses sound in overlapping bands often described as critical bands. Within a band, the ear largely judges energy together rather than separating every individual tone. So when one instrument dominates a band, anything else competing in that same band gets pushed under its threshold.
Two kinds of masking matter when you mix:
- Simultaneous (spectral) masking — a loud sound hides a quieter sound playing at the same moment in the same frequency region. This is the everyday culprit in dense mixes.
- Temporal masking — a loud sound can hide quieter sounds immediately before and after it in time, not just during it.
Where it shows up in real mixes
Masking gets worse the more sources you stack into the same range. Common collision zones:
- Low-mids (roughly 200–500 Hz): bass, low strings, piano left hand, kick body and male vocals all pile up here. The result is boxy, congested mixes.
- Vocal range (roughly 1–4 kHz): lead vocals fight guitars, synths and upper strings for presence and intelligibility.
- Orchestral and acoustic material: a Decca tree, spot mics and a choir capture overlapping content of the same instruments — natural, beautiful, and densely masked.
How engineers usually fight masking
The classic tools all work, but each has a cost:
- Static EQ: carve a dip in one track to make room for another. Effective but blunt — it cuts even when the two instruments aren't actually colliding, and it only knows about the one track it's on.
- Arrangement and panning: separate sources in time or in the stereo field so they collide less. The best fix when it's available — but you can't always re-write the part.
- Dynamic EQ and sidechaining: duck one track only when another plays. Powerful, but it has to be set up per pair of tracks, by hand, with routing.
The common limitation: a plug-in on a single track can only see that track. It has no picture of what the rest of the session is doing at any given moment.
De-masking: fixing it across the whole mix
De-masking tackles the problem directly: instead of guessing with a fixed EQ curve, it carves frequencies only where and when instruments actually compete, and only by as much as needed. Done with minimum-phase processing, it makes room without the time-smearing or "underwater" artifacts that linear-phase or FFT-based approaches can introduce.
The next step is cross-track de-masking: rather than one instance cleaning one source in isolation, every instance shares a picture of the whole project and they make room for each other automatically — no sidechain routing, no manual pairing. That's the approach Spectral Engine takes, and it's built for exactly the dense, organic material — orchestra, choir, acoustic ensembles — where masking is hardest to tame by hand.
Spectral Engine clears frequency masking across your whole project, automatically. Join the beta for early access.
Join the beta