Phase distortion is a change in the timing relationship between different frequency components of a signal as it passes through a system. Every signal, whether it’s sound, light, or digital data, is made up of multiple frequencies. When those frequencies travel through a filter, speaker, lens, or any other medium, they can be delayed by different amounts. If all frequencies are delayed equally, the signal comes out intact, just a little later. If some frequencies are delayed more than others, the shape of the signal changes. That change is phase distortion.
How Phase Distortion Works
Any filter or transmission system introduces some delay between its input and output. In the time domain, this is simply a transmission delay. In the frequency domain, that same delay shows up as a phase shift. When the phase shift is perfectly proportional to frequency (called a linear phase response), every component of the signal is delayed by the same amount of time. The output is an exact, slightly delayed copy of the input. No distortion occurs.
Phase distortion happens when that proportionality breaks down. With a nonlinear phase response, a low-frequency component might be delayed by one amount while a high-frequency component is delayed by a different amount. The components arrive at the output at slightly different times relative to each other, and when they recombine, the signal’s shape is altered. Unlike simple amplitude distortion, which changes how loud certain frequencies are, phase distortion rearranges when those frequencies arrive. The result can be subtle or dramatic depending on the system.
Group Delay and Phase Delay
Engineers use two related measurements to describe how a system affects phase. Phase delay tells you how much a single pure tone is delayed as it passes through a filter. Group delay tells you how much a cluster of nearby frequencies, like those making up a note or a pulse, is delayed as a group. For a system with perfectly linear phase, these two numbers are identical and constant across all frequencies. That’s the ideal scenario: everything passes through at the same speed.
When the phase response is nonlinear, group delay varies with frequency. Some frequency bands pass through quickly while others lag behind. This frequency-dependent delay is the mechanism behind phase distortion. In practical terms, a narrow-bandwidth signal (like a single musical note) is delayed by the group delay at its frequency, while the fine structure of the wave within that note is delayed by the phase delay. When these don’t match, you get a smeared, altered version of the original.
Phase Distortion in Audio
In audio systems, phase distortion most commonly arises in crossover networks, equalizers, and loudspeakers. Every filter, whether active or passive, alters phase. A crossover network splits an audio signal into frequency bands for different drivers (tweeter, woofer), and the phase shifts introduced by those filters can change how the bands recombine at your ears.
Of all standard crossover filter types, only the simplest 6 dB per octave design preserves both accurate amplitude and phase when the outputs are reassembled. More aggressive crossover slopes provide better frequency separation but introduce greater phase shifts. Some filter pairs are “phase complementary,” meaning their combined output restores the original phase relationships, but many commonly used crossover designs are not.
Whether listeners can actually hear phase distortion is a long-running debate. The human ear appears unable to detect phase shifts in steady, continuous tones. But phase matters for transient sounds, like drum hits or consonant sounds in speech, where the precise alignment of frequency components defines the sharp attack. Changes in phase between harmonically related components of a complex sound can alter the perception of both timbre and pitch. As a rough guideline, phase changes of less than 90 degrees per octave are generally inaudible. For hearing aids and similar devices, phase distortion at speech frequencies is considered negligible. Still, as other forms of distortion in high-fidelity systems have been reduced over the decades, phase distortion has become more noticeable by comparison, particularly in stereo imaging, where phase cues help your brain determine the direction sounds are coming from.
Linear Phase vs. Minimum Phase Filters
In digital audio and signal processing, two filter designs represent opposite trade-offs around phase distortion. Linear phase filters (typically FIR filters) delay all frequencies by exactly the same amount, eliminating phase distortion entirely. The catch is latency: the more aggressive the filter, the greater the delay through it. Linear phase filters also produce pre-ringing, a faint echo that appears before a transient rather than after it. This artifact is unavoidable with linear phase designs and can sound unnatural on percussive material.
Minimum phase filters (typically IIR filters) have lower overall latency and no pre-ringing, but they introduce frequency-dependent group delay. Some frequency bands pass through faster than others. They do produce post-ringing, but this tends to be masked by the signal itself and is less perceptible than pre-ringing. Audio engineers choose between these designs depending on context. Linear phase equalizers are common in mastering, where preserving the exact shape of transients across a full mix matters. Minimum phase designs are preferred for live sound or tracking, where low latency is essential.
Phase Distortion in Optics
Phase distortion isn’t limited to sound. In optical systems, light waves passing through lenses, windows, or the atmosphere can experience uneven delays across the wavefront. When light passes through glass, variations in the refractive index caused by heat or mechanical stress create nonuniform delays. Different parts of the wavefront arrive at the sensor or your eye at slightly different times, corrupting the image.
Atmospheric turbulence is another major source. The shimmering you see above hot pavement, or the twinkling of stars, is phase distortion at work. For precision systems like telescopes, tracking systems, and directed-energy devices, these phase aberrations cause blurring, pointing errors, and focus problems. Adaptive optics systems correct for this by measuring the distorted wavefront in real time and adjusting a deformable mirror to cancel out the phase errors.
Phase Distortion as a Synthesis Technique
In music synthesis, phase distortion is not just a problem to solve but a creative tool. Casio popularized phase distortion synthesis in the 1980s as an alternative to the more expensive FM synthesis of the era. The basic idea is clever: instead of reading through a stored waveform (like a sine wave) at a steady rate, you distort the rate at which you read through it.
In standard wavetable synthesis, a smooth, linear ramp signal acts as a pointer that moves evenly through one cycle of a waveform. Phase distortion synthesis bends that ramp, putting a kink or knee at a designated point. The pointer speeds up through part of the waveform and slows down through the rest. When this distorted ramp reads through a sine wave, the output is a reshaped waveform with richer harmonic content. The sharper the bend, the more harmonics appear. By sweeping the kink point over time, you can create evolving timbres that mimic the brightness changes of acoustic instruments, all from a single sine wave lookup table.
Correcting Phase Distortion
When phase distortion is unwanted, all-pass filters are the primary correction tool. An all-pass filter passes every frequency at equal amplitude but shifts their phase. By carefully designing an all-pass filter with the opposite phase response of the distortion you want to fix, you can realign the frequency components without changing the signal’s tonal balance. This technique is used in audio crossover design, digital signal processing, and even spectroscopy, where phase errors in measured data can be canceled to very high precision using low-order digital all-pass filters.
In digital audio workstations, some plugin designers include phase correction stages after processing. In loudspeaker design, time-aligning the drivers physically (stepping the tweeter back relative to the woofer, for example) addresses phase differences caused by the crossover. Each approach targets the same underlying problem: making sure all the frequency components of a signal arrive at their destination at the right time relative to each other.