Linear Phase vs Minimum Phase EQ: When to Use Each

What is the difference between linear phase and minimum phase EQ?

A minimum phase EQ shifts the timing of frequencies near the band you adjust, so boosted or cut frequencies arrive slightly later than the rest. A linear phase EQ delays every frequency by the same amount, so it changes tone without changing timing relationships inside the sound.

iZotope describes linear phase EQ as avoiding the phase distortion of minimum phase EQ by delaying all frequencies equally^[1]. FabFilter puts it as linear-phase filters changing only the magnitude of the audio while leaving the phase untouched^[2]. Neither is better in general. They trade different side effects.

When can you actually hear EQ phase shift?

On a single track you usually cannot. Phase shift becomes audible when the EQ'd signal is combined with another copy of the same sound, because the two copies no longer line up at the frequencies you changed.

FabFilter notes that phase effects from a simple bell or shelf are very subtle, while steep low-cut and high-cut filters can start to affect transients^[2]. iZotope's example is a snare with mics above and below: EQ one mic with a minimum phase EQ and the existing phase relationship between them changes^[1]. Image-Line gives a producer's version: phase rotation mainly causes artifacts when mixed with other channels, such as losing bass from a Kick^[4].

If phase itself is still an abstract idea, what is phase rotation and phase cancellation explained cover the basics with drum examples.

What is pre-ringing in a linear phase EQ?

Pre-ringing is a faint smear of filtered energy that starts before a transient. Because a linear phase filter spreads its response evenly before and after each moment, a sharp hit gets a soft swell leading into it.

FabFilter warns that higher linear-phase resolution can introduce pre-ring that makes transients, such as a kick drum, lose their edge^[2]. Apple says the same about Logic's Linear Phase EQ: it affects the onset of the transient, most visibly with steep cut filters or high boosts and cuts on narrow bands^[3].

1. Solo a punchy kick
   Pick a dry kick one-shot with a sharp click at the start.
2. Add a steep low cut
   Set a 48 dB/oct high-pass around 40-50 Hz, or a narrow 6 dB cut near 60 Hz.
3. Switch the mode
   Toggle between minimum phase and linear phase at the highest resolution you have.
4. Listen before the hit
   In linear phase, a soft woof or swell may appear just before the click. That is pre-ringing.
5. Reduce it
   Use a gentler slope such as 12-24 dB/oct, a wider Q, a lower resolution, or go back to minimum phase.

How much latency and CPU does linear phase EQ cost?

A lot more than minimum phase EQ. Linear phase filters buffer audio before processing, and better low-frequency accuracy needs a longer buffer, so latency rises with resolution.

FabFilter publishes the numbers for Pro-Q's Linear Phase mode at 44.1 kHz^[2]. Using L/R and M/S bands at the same time doubles that latency, because the processing runs in two stages.

Your DAW hides this delay during playback with plugin delay compensation, which Image-Line notes for FL Studio's EQ^[4]. It cannot hide it while you play or record live. See plugin delay compensation explained for when that alignment breaks.

When should you use linear phase EQ?

Use linear phase EQ when an EQ'd signal will be summed with another version of itself, or when you need to split and recombine frequency bands cleanly. Outside those cases, it is rarely worth the latency.

• Multi-mic and layered sources Apple recommends Logic's Linear Phase EQ where phase coherence between tracks is needed, such as multi-microphone recordings^[3]. Snare top and bottom, kick in and out, and doubled 808 layers all qualify.
• Parallel processing If an EQ'd copy is blended with the dry track, FabFilter notes that mismatched phase makes the copies partly cancel^[2]. Linear phase keeps setups like parallel EQ on drums predictable.
• Band splitting FL Studio's Frequency Splitter offers zero-latency minimum phase crossovers or linear phase crossovers for perfect recombination with the dry signal, with slopes from 6 to 96 dB/oct^[5].
• Partial master EQ FabFilter's mastering example: EQ applied to one section with crossfades will not blend smoothly if the filtered and unfiltered parts differ in phase^[2].
• Mid/side work on a master Separate mid and side EQ curves can shift the two channels differently. Linear phase avoids that; the trade-offs are in mid-side EQ on masters.

When is minimum phase EQ the better choice?

Most of the time. FabFilter states that in almost all normal mixing and mastering situations its Zero Latency or Natural Phase modes are the best choice, and that linear-phase processing is not better or more transparent than normal processing, it is different^[2].

Kick, 808 and the low end

Better low-frequency resolution needs longer linear phase processing^[2], so pre-ringing is most likely exactly where trap and drill mixes live. For a kick or 808 on its own track, use a minimum phase high-pass at 12-24 dB/oct as a starting point and keep the transient intact. If the kick and 808 cancel each other, fix their alignment and tuning first; phase issues between 808 and sub walks through it.

Tracking, live play and character EQ

Latency rules linear phase out while recording or playing a MIDI keyboard. Minimum phase is also the right pick when you want an EQ to sound like hardware, since analog filters shift phase as well^[2] and that behavior is part of their tone.

Linear phase and minimum phase modes in common EQs

Many modern EQs offer both behaviors behind a switch. Knowing the label saves you hunting through menus.

When you compare modes, match output levels, loop a section with clear transients, and listen to the attack of the drums rather than the overall tone. If you cannot hear a difference, keep the zero-latency mode.