Masking is one of the most important concepts in mixing music. It describes how one sound can obscure another, making it difficult—or even impossible—to hear. You’ve likely experienced this both in everyday life and in your own production work.
Masking doesn’t just occur with broadband noise like white noise; even a single resonant peak can make certain frequencies harder to perceive. By understanding how sounds interfere with each other, we can take steps to prevent this in our mixes.
Basic Examples of Masking
To illustrate masking, let’s start with some simple audio demos using sine waves and noise. These basic examples will help us understand the concept before we apply it to real mixing scenarios.
White Noise Masking a Tone
In this first demo, I’ll play a pure tone, then gradually increase the level of white noise until the tone is completely masked. Even though the tone remains at a constant volume, you’ll hear it disappear as the noise level increases. When I turn the noise off, the tone will suddenly “reappear.”
Narrowband Masking
Most elements in a mix aren’t full-spectrum noise, but they can still mask each other. In this next example, I’ll filter the white noise down to a narrow band and then introduce a sine wave close to that frequency range. As I fade the noise in and out, listen to how the sine wave gets masked.
Additionally, if I move the sine wave up and down in frequency, you’ll notice how it becomes masked when it enters the same frequency range as the noise and re-emerges when it leaves.
Critical Bands and Acoustic Shadows
Luckily for us as music producers, masking only affects frequencies close to each other. Our ears can focus on specific frequency areas, which are called critical bands. If two sounds fall within the same critical band, they are at risk of masking each other.
To illustrate this, I’ll mask a tone using pink noise but then apply a notch filter to remove the noise around that tone. As the noise moves outside the critical band, the tone will become audible again. The moment the notch becomes narrower, the noise reenters the critical band, making the tone harder to hear.
These auditory filters are roughly one-third of an octave wide, though they slightly widen in the lower frequencies. A helpful way to visualize this is as an acoustic shadow—any sound that falls under this shadow is likely to be masked.
Upward Spread of Masking and the Muddy Low-Mids
One particularly important aspect of masking is the upward spread of masking—the louder a sound is, the more high frequencies it masks above it. This effect is nonlinear; while lower frequencies are affected to a lesser extent, frequencies above the masker get disproportionately more obscured.
This is why the low-mid frequencies (200Hz–500Hz) are notorious for making mixes sound muddy—when these frequencies get too loud, they start interfering with the clarity of instruments in higher ranges.
A key takeaway here is to check your mix at different playback volumes. What sounds balanced at low volume might become masked as you turn it up.
Techniques to Reduce Masking
1. Arrangement & Sound Design
The most effective way to prevent masking isn’t through mixing—it’s through arrangement, songwriting, and sound design.
- Avoid placing instruments in the same frequency range if you want them to be clearly audible.
- If two melodic instruments are competing, shift one up or down an octave to separate them in the spectrum.
- Keep the vocal range clear by minimizing overlapping instruments.
2. EQ Unmasking
If arrangement isn’t an option, EQ is your best tool for unmasking sounds.
- Identify which element is more important in a given frequency range.
- Use subtractive EQ to dip the masking frequencies in the background element.
- A common approach is high-passing all instruments except for the kick and sub-bass, ensuring clarity in the low end.
3. Dynamic EQ & Sidechain Unmasking
A limitation of static EQ is that instruments don’t stay in one place—they change notes, come and go, and occupy different frequency bands at different times.
To address this, dynamic EQ or sidechain multiband compression can help. These tools dynamically reduce masking frequencies only when necessary.
In this demo, I have a piano melody and a noise pad. The piano moves in pitch, making static EQ adjustments ineffective. Instead, I set the piano as a sidechain input to a spectral unmasker, which selectively ducks the frequencies of the noise when the piano plays. Listen to how the piano suddenly “lifts” out of the background.
For a simpler approach, broadband sidechain ducking can work when two full-spectrum elements compete. For example, a snare and an instrument bus—ducking the instrument bus when the snare hits makes the snare stand out more clearly.
Creative Uses of Masking
While masking is usually something to avoid, it can be used creatively.
For example, in my track “Rats, Bats, and Critters”, I deliberately mixed the vocal quietly against a noisy beat, making it hard to hear. This created a mysterious, distant “muttering” effect instead of a lead vocal presence.
Masking in MP3 Compression
Masking is also heavily used in psychoacoustic audio compression (e.g., MP3). The goal here isn’t to fix masking but to exploit it to remove inaudible information and reduce file size.
MP3 encoding analyzes the sound spectrum and discards:
- Sounds below the human hearing threshold (based on Fletcher-Munson curves).
- Sounds fully masked by other elements.
Using a plugin like Unmask, we can listen to the audio data an MP3 encoder would remove. Here’s a demo where I play a song normally, then isolate the “masked” information that would be discarded.
Temporal Masking
So far, we’ve only discussed frequency masking, but masking also occurs over time.
Forward Masking
A loud sound can mask a quiet sound that follows immediately after. For instance, if I play a quiet tone, you can hear it clearly. But if I precede it with a loud sound, the quiet tone becomes masked—even though they don’t overlap in time.
Backward Masking
Strangely, masking can also occur backwards in time. If a quiet sound is immediately followed by a much louder sound, our brain essentially “forgets” that it heard the quieter one.
In mixing, this is most relevant for percussion and transients. This is why sidechain compression with lookahead can help preserve transients. In our snare ducking example, adding lookahead makes the transient sound much snappier.
Stereo Masking
Until now, we’ve assumed everything is mono, but stereo imaging plays a huge role in masking.
- Sounds are easier to hear if panned away from competing elements.
- Delaying one side of a sound can unmask it in stereo playback.
However, these techniques don’t always translate well to mono systems, so be mindful of your audience.
