Introduction

Equalizers are the primary toolkit for shaping the tonal balance of audio. Most engineers can quickly identify a 3 dB boost at 5 kHz or a high-pass filter at 80 Hz. Yet the debate between linear phase and non-linear phase (minimum phase) EQ remains one of the most nuanced in audio engineering. The choice directly impacts transient response, stereo imaging, and the overall "feel" of the mix.

This guide breaks down the engineering principles of phase equalizers. It moves past simple definitions to explore the physics of phase shift, the engineering trade-offs of FIR and IIR filters, and the practical workflow decisions that professional engineers make every day. Understanding these concepts allows you to reach for the right tool with confidence, whether you are tracking a live band, mixing a dense pop track, or mastering an audiophile recording.

The Physics of Phase in Audio

Phase describes the position of a waveform in time. A full cycle of a waveform is 360 degrees. When an EQ applies a filter, it does not just change the amplitude of a frequency; it also shifts the timing of the frequencies around the center point. This is phase shift.

Standard analog-style equalizers introduce frequency-dependent phase shifts. For example, a low-pass filter shifts the phase of the signal by -90 degrees at the cutoff frequency. This shift is an unavoidable byproduct of the filter's design. The key difference between linear and non-linear phase EQs is how they handle this timing relationship.

Constructive and Destructive Interference

Phase shift becomes critical when multiple microphones capture the same source, or when parallel processing is used. If two versions of a signal are summed, one with a phase shift and one without, they can cancel each other out at specific frequencies. A non-linear phase EQ can subtly alter the phase relationship within a single track, affecting how it sits in the mix. Linear phase EQs are often chosen for mastering because they preserve the original phase alignment of the stereo mix, preventing the EQ itself from disturbing the carefully balanced phase relationships between instruments.

Group Delay vs. Phase Shift

It is important to distinguish between phase shift (measured in degrees) and group delay (measured in milliseconds). Group delay is the actual time delay experienced by a specific group of frequencies. A minimum phase EQ delays the frequencies near the filter cutoff. The higher the Q and the higher the gain, the longer the group delay. A linear phase EQ applies a constant group delay to all frequencies. This preserves the waveform shape but introduces significant latency.

Non-Linear Phase (Minimum Phase) Equalizers

This category encompasses the vast majority of classic analog hardware and digital emulations. Pultec, Neve, API, SSL, and the standard digital EQ in your DAW (like Ableton Live's EQ Eight or Logic's Channel EQ) are minimum phase.

Why "Minimum Phase"?

The term "minimum phase" has a specific mathematical meaning. For the magnitude response (the boost or cut curve) that you dial in, the filter produces the minimum possible delay. It concentrates the filter's energy right at the transient event. This is why analog-modeled EQs feel "snappy" and immediate. The transient hits, the filter responds, and the phase shift occurs naturally afterwards.

The "non-linear" aspect refers to the relationship between frequency and phase shift. A 3 dB boost at 100 Hz causes a different phase rotation than a 3 dB boost at 10 kHz. This varying phase response is what gives minimum phase EQs their character.

Sound Characteristics and Musicality

Minimum phase EQs are often described as "musical." This is because our ears are accustomed to this type of phase distortion. Acoustic instruments and analog circuitry naturally produce phase shifts. When you boost a snare drum with a minimum peak EQ, the slight delay around the center frequency can add thickness and weight that sounds natural.

  • Strengths: Low latency, natural sounding transients, efficient CPU usage, excellent for adding character.
  • Weaknesses: Can smear transients when used aggressively, phase shifts can muddy low-end if multiple cuts are stacked.
  • Best Use Cases: Tracking, live sound, mixing individual tracks, creative sound design, emulating analog hardware.

Linear Phase Equalizers

Linear phase EQs solve the problem of frequency-dependent group delay. They ensure that all frequencies pass through the filter with the same time delay. This eliminates the phase shift associated with the EQ curve.

How They Work: FIR Filters

To achieve this, linear phase EQs use Finite Impulse Response (FIR) filters. Instead of using feedback (which creates phase shift), FIR filters use a tapped delay line. The filter "looks" at a block of audio samples, calculates the required output, and applies the EQ. This is a convolution process. Because the filter requires a full block of audio to make its calculation, it introduces latency.

The impulse response of a linear phase filter is symmetrical. This symmetry is the source of both its phase coherence and its most controversial artifact: pre-ringing.

The Pre-Ringing Artifact

Because the filter is symmetrical, energy is generated before the transient event. If you have a sharp transient, like a kick drum hit, a high-Q linear phase boost will cause a "swoosh" or "knock" sound that occurs a few milliseconds before the actual hit. This is highly unnatural to the human ear. We expect filters to react after a transient, not before.

Pre-ringing is most audible with high gain, high Q settings. A gentle shelf boost or a low-Q cut is much less likely to cause audible pre-ringing. This is why linear phase EQs are often preferred for broad mastering adjustments, but can sound terrible on percussive elements like drums or plucked strings.

Latency and CPU Load

Linear phase EQs require significantly more processing power and introduce substantial latency. An FIR filter for a linear phase EQ might need thousands of taps (samples) to achieve its response. This makes them unsuitable for live sound or tracking. In a DAW, automatic delay compensation handles the latency, but it can add up quickly if multiple linear phase EQs are used across a session, causing lag during playback.

Choosing the Right EQ for the Task

There is no "better" type of EQ. The choice depends entirely on the material and the goal.

Mastering and Mix Bus

Linear phase EQs are the standard for mastering. The goal is to shape the tonal balance of the final mix without introducing artifacts that degrade the stereo image or phase coherence. Broad, gentle curves with linear phase EQs can add air or clarity without smearing the master bus.

Many mastering engineers use a hybrid approach. They might use a minimum phase EQ for a subtle low-end shelf (where the natural phase shift can help tighten the bass) and a linear phase EQ for a high-frequency air band (where preserving transient detail is critical).

Tracking and Mixing

For tracking and mixing individual elements, minimum phase EQs are almost always the better choice. The low latency allows for real-time monitoring. The natural transient response keeps drums punchy and vocals immediate. If you need to make a surgical cut to remove a resonance, a minimum phase EQ with a high Q will work well, though you should listen for "ringing."

There are specific mixing scenarios where linear phase EQs shine:

  • Canceling Room Modes: Cutting a standing wave frequency with a minimum phase EQ can smear the transient of the kick drum. A linear phase cut can be cleaner.
  • Parallel Processing: Using a linear phase EQ on the parallel chain ensures that the phase relationship with the dry signal remains coherent.
  • Mid/Side EQ: Linear phase EQs are often preferred for mid-side processing because any phase shift between the mid and side channels can cause the stereo image to collapse or become unstable.

Advanced Workflow Considerations

Professional engineers often use a combination of both types within a single session. The key is to understand the trade-offs.

Cascading Filters

Stacking multiple minimum phase EQs can lead to significant group delay accumulation. If you have five cuts spread across five different minimum phase EQs on a single channel, the transient response can become noticeably soft. In this case, it is often better to use one instance of a linear phase EQ to handle the bulk of the corrective cuts, or to use a single, more complex EQ plugin.

Phase Correlation Metering

Use a phase correlation meter on your master bus. If you engage a linear phase EQ and the correlation meter shifts, it indicates that the EQ is changing the phase relationship. This is a good thing if it improves clarity, but a warning sign if it causes the low end to lose power. A minimum phase low-cut filter can often increase the apparent punch of a kick drum because of the phase shift it introduces. A linear phase low-cut filter will be more transparent but may feel less impactful.

Hybrid Mode (Natural Phase)

Some modern EQs, like the FabFilter Pro-Q 3, offer a "Natural Phase" mode. This attempts to balance the latency and pre-ringing of linear phase with the natural response of minimum phase. It is a practical compromise for many mixing scenarios. These hybrid modes are worth experimenting with as they often provide the "best of both worlds" for general mixing.

Conclusion

The difference between linear and non-linear phase EQs is a fundamental choice in audio processing. Minimum phase EQs offer the natural, musical response we expect from classic hardware, with low latency and efficient CPU usage. Linear phase EQs offer absolute phase coherence, making them indispensable for mastering and critical surgical tasks. By understanding the engineering principles of group delay, pre-ringing, and filter design, you can select the right tool for each specific task. Trust your ears, but verify your phase correlation, and do not be afraid to switch between modes to find what serves the music best.