The ultimate headphone EQ guide for
mixing & mastering.

If you mix or master on headphones and you are not EQ'ing them, you are working blind. That is not an opinion — it is the single most important variable standing between you and mixes that translate.

This guide pulls together the real-world methodology of Emrah Celik, an award-winning mix and mastering engineer with over three billion streams (90% mixed on headphones), and Paul Third, co-host of MixPhones. What follows is not theory plucked from a forum — it is a battle-tested workflow refined over years of professional headphone mixing.

We live in a consumer world. Headphones are designed to sound exciting to buyers, not accurate to engineers. Even models marketed as "studio" headphones almost always ship with a hi-fi curve — a baked-in tonal character that flatters music playback rather than revealing it honestly.

If every headphone sounded the same, there would be no market for hundreds of competing models. Each one has its own signature: boosted bass here, scooped mids there, a peak in the treble. That signature is what makes your mixing decisions unreliable out of the box.

As Emrah puts it plainly:

"Every headphone has huge frequency deviations — sometimes 5 to 10 dB. Maybe 10 dB bright, ear-piercing bright, or there's no bass, like 5 dB lost on your tonality. EQ doesn't colour your sound. It corrects those flaws so you are hearing an accurate reference."

Without correction, you counter-mix against whatever the headphone is doing wrong. If your cans are bass-light, your mixes come out bass-heavy. If they are too bright, your mixes end up dull. Paul recalls his early days before EQ correction:

"My headphones were really muddy sounding, so my counter translation was that I used to always make my mixes sound too thin. I'd send them to mastering engineers and they'd politely tell me it doesn't translate. Once I started EQ'ing to a target, for the first time ever my translation was really good to speakers."

The bottom line: mixing without EQ'ing your headphones is like colour grading a film on an uncalibrated monitor. You cannot trust what you hear.

The Harman target curve is the result of extensive scientific research into how headphones should be tuned so that they are tonally akin to speakers in a well-treated room. It is not an arbitrary preference — it is grounded in psychoacoustic research led by Dr. Sean Olive and the team at Harman International.

There is a common objection online: "Harman is just a preference target — why would mixing engineers use a preference target?" Paul and Emrah address this head-on:

"Harman has been tested. As soon as I made the video talking about why I advise using the Harman curve for mixing, so many people started using it and they were talking about how good their translation was. The only element of preference is really in the low end." — Paul

The scientific purpose of Harman is to recreate the tonal balance of calibrated studio monitors in a treated room. The low-end shelf is the part that varies most between listeners — some people are more sensitive to bass than others. But the mid-range and upper-mid response is well-researched and consistent.

Known Characteristics of Harman

• Slightly bright: both Paul and Emrah note that Harman can lean bright. You may want to tame the upper mids slightly for mixing work.
• Full low-end shelf: the bass shelf is generous. If your mixes translate too bass-heavy, pull this back by a dB or so.
• Low-mid dip: Harman has a dip around 300–360 Hz that can sometimes cause you to let too much low-mid energy through. Emrah recommends adding a small bump at 360 Hz to compensate: "Just show me a song that you take down 360 Hz a few dB and it didn't sound good."

The key insight: use Harman as your starting point, not your finish line. It will get you 85–90% of the way. The remaining 10–15% comes from learning how your particular EQ translates and refining with references over time.

This is the section that trips up even experienced engineers. The instinct is to reach for a linear phase EQ — the reasoning being "I just want to change the tonal balance without affecting the phase." It sounds logical. It is wrong for headphone correction.

Emrah explains the science concisely:

"There is no linear phase in the analog domain. Headphones are analog devices. When they have a frequency response, they also have a phase response. If you don't change the phase and keep it linear, you break the sound. You have to fix the phase as well. The only way to fix phase is to use minimum phase, not linear phase — it's counter-EQ'ing inclusive of phase."

Headphones are acoustic transducers. Their frequency response deviations come with corresponding phase shifts. A minimum phase EQ naturally applies the inverse phase correction alongside the frequency correction. A linear phase EQ only corrects the frequency response and leaves the phase errors intact — which degrades imaging and transient accuracy.

The Rule

You do not need FabFilter Pro-Q3 or any boutique plugin. Stock digital EQs in Logic, Ableton, Reaper, or any other DAW will do the job perfectly. As Paul notes: "A linear EQ is a linear EQ — they will null. It's ones and zeros."

A Note on Pro-Q3

If you do use FabFilter Pro-Q3, be aware that it uses biquad filters with a different Q convention. You need to multiply the Q value by 1.41 to match what other EQs expect. For example, if the target Q is 1.0, set Pro-Q3 to 1.41. Or simply use a stock EQ and avoid the hassle entirely.

Sonarworks (now dSONIQ SoundID Reference) is the most well-known headphone correction software on the market. It is also, in the view of Paul and Emrah, fundamentally flawed in its approach for mixing purposes.

The problem: Sonarworks uses mirror correction. It measures the frequency response of your headphone and applies a correction curve that perfectly inverses every deviation from the target — every tiny dip and peak, corrected with narrow, precisely-placed bands.

Why This Is a Problem

• Phase distortion: when you stack many narrow EQ bands close together, you introduce phase distortion that smears your stereo image and degrades transient accuracy.
• Destroys headphone character: the small peaks and dips in a headphone's response are part of its acoustic character. Correcting every one of them removes the natural sound of the driver.
• Over-correction: Sonarworks may use 20 or 30 bands internally. You have no control over how many bands are being applied or how narrow they are.

As Emrah explains:

"Those small variations in the frequency response are the character of the headphones that you don't want to mess up. It doesn't have to perfectly follow the target — just connect the main dots without chasing every dip and peak."

The MixPhones approach is the opposite of Sonarworks: use the fewest bands possible to get close to the target, preserve the headphone's natural character, and avoid phase distortion.

Paul's recommended workflow has evolved over time. He previously used Auto EQ but now advocates for manual EQ using Squiglink, the web-based frequency response comparison tool. Here is the full process:

1. Get Your Headphone's Frequency Response

Go to Squiglink and search for your headphone. If it is measured by Oratory1990 (the gold standard for consistency, using a GRAS measurement rig), use that measurement. If it is not on Squiglink, you can download a CSV file from the Auto EQ GitHub repository and upload it to Squiglink.

2. Set Your Target

Select the Harman target as your baseline. Squiglink lets you overlay multiple targets and measurements, so you can visually compare your headphone's response against the target.

3. Identify the Key Corrections (Fewest Bands Possible)

Look at where your headphone deviates most from the target in the range of 30 Hz to 6 kHz. This is the critical region. Do not try to correct above 10 kHz — those peaks and dips are caused by the interaction between the headphone driver, your ear canal, and your pinna. They are supposed to be there.

4. Apply Your EQ

Emrah's recommended band structure for most headphones:

"The sweet spot is around four to five bands. More than that is unnecessary. If you pick the right headphones, a low shelf, a high shelf, and two or three mid-range bands is enough." — Emrah

5. Use Wide Q Values

Do not chase every little peak with narrow notch filters. If there is brightness between 3–6 kHz, use a single wide band to gently bring it down rather than targeting individual spikes. Narrow notches close together cause phase distortion.

6. Do NOT EQ Above 10 kHz

The region above 10 kHz shows dramatic peaks and dips on every measurement. These are caused by ear canal resonances and pinna reflections — they are part of how you hear, not flaws in the headphone. Trying to flatten this region will make things sound worse, not better.

7. Learn the Sound, Then Refine

Once your EQ is set, mix with it. Use reference tracks. Compare your headphones against speakers in a treated room if you can. Let your translation results tell you what to tweak — that is your personal HRTF in practice.

EQ is only half the equation. If you are listening at the wrong level, your headphones will mislead you regardless of how well they are corrected.

The science behind this is the Fletcher-Munson curve (equal-loudness contours). At low listening levels, human hearing is dramatically less sensitive to bass and treble. At around 85 dB SPL, our hearing flattens out — we perceive bass, mids, and highs in a much more balanced way.

Why 85 dB SPL?

"If you look at the Fletcher-Munson curve, 85 dB SPL is where our ears are mostly flat — similar reaction to bass, mids, and highs. That's why you have to mix and master on headphones at 85 dB. All the details open up and it makes you less heavy-handed — you use less distortion, less compression." — Emrah

At 70–75 dB SPL, many engineers complain that their planar headphones lack slam and bass impact. The headphones are not the problem — the level is. Crank to 85 dB SPL and the headphones transform. The low end fills in, transients become punchy, and micro-details emerge.

Will 85 dB SPL Damage My Hearing?

No. Research shows that 85 dB SPL is safe for approximately 8 hours of continuous exposure. For context, driving a car typically exposes you to 75–80 dB, and living on a busy street can reach 90–95 dB. You are not going to damage your hearing at 85 dB SPL.

How to Measure Your Headphone Level

You have two options:

• SPL meter: place it between the ear cups with your music playing. Set it to slow response and either dBA or dBC weighting.
• Phone app: not scientifically perfect, but it will get you in the ballpark.

dBA vs dBC Weighting

Emrah also recommends using brown noise (not pink, not white) to calibrate your listening level. Brown noise has a tilt that more closely represents the energy distribution of real music. Play it, set your volume, mark that position, and keep it consistent.

"You don't need to EQ headphones"

This is a dead giveaway that someone does not mix professionally on headphones. As Paul says: "If you ever want to know if somebody actually mixes professionally on headphones and they say 'I don't EQ my headphones' — most likely they don't mix on headphones." There is no perfect headphone on the market. Even the Dan Clark Audio Noire RX, which Paul expected might not need correction, required five bands of EQ.

"Low impedance headphones need less power"

This is a widespread misconception rooted in 20-year-old thinking about Beyerdynamic and Sennheiser headphones. Impedance determines how power is delivered (voltage vs current), not how much is needed. The real factor is sensitivity. A 32-ohm planar magnetic headphone can be far harder to drive than a 300-ohm dynamic because planars demand high current. Most audio interfaces are severely current-limited.

"A MacBook Pro is enough — you don't need an amp"

The MacBook Pro has arguably the best headphone output of any laptop. It will get many headphones to adequate volume. But volume is not accuracy. As Emrah states: "You get sound from a MacBook, but not accuracy. The moment you get into professional or high-end headphones, you need a headphone amp." Even the Slate VSX, which many assume is easy to drive, reveals dramatically more detail through a proper amp. You do not need to spend a fortune — a Topping DX1 at around 99 GBP is enough.

"Auto EQ convolutions are the best correction"

Auto EQ algorithms chase perfection. They create corrections with potentially 20–30 bands to perfectly match a target line. This over-correction introduces phase distortion and removes the natural character of your headphones. If you use Auto EQ, Paul recommends limiting the max Q to about 3, restricting the frequency range to 20 Hz – 8 kHz, and treating it as a starting point only. Better yet, skip it entirely and EQ manually using Squiglink.

"Listening quietly protects your hearing and is fine for mixing"

Listening too quietly is one of the most common mistakes. At low levels, you simply cannot perceive bass accurately due to the Fletcher-Munson effect. Your headphones sound different at different volumes — it is not a linear experience. If you mix at 70 dB SPL, you will over-compensate on bass and under-compensate on dynamic range. Get to 85 dB SPL and let your headphones breathe.

For those who want to push headphone correction further, Emrah and Paul have been testing convolution-based EQ and found measurable differences compared to standard parametric EQ.

Using bespoke AB testing software, Emrah measured the difference between the same correction applied via parametric EQ versus a convolution (impulse response). The result: 100% detectable difference in blind testing, primarily in transient response and high-frequency speed.

"The response is faster compared to EQ. The post-EQ ringing is less than normal EQ. That's the good part of using convolution." — Emrah

You can create an impulse response from your parametric EQ and load it into MConvolutionEZ (a free plugin by MeldaProduction). The same rule applies: keep the correction minimal. A convolution with fewer corrections will outperform one that over-corrects.

Emrah is also researching neural amp modelling for headphone correction — a potential future development that could change the game entirely.

EQ & Measurement

System-Wide EQ

Hardware Options

Plugin-Based EQ

Not all headphones respond well to EQ. If a headphone does not produce a frequency range naturally, no amount of boosting will create it cleanly. Here is what to look for and how to test:

What Makes a Good Candidate

• Full bandwidth: the headphone should extend cleanly into the sub-bass. If it rolls off steeply below 80 Hz (like the HD650), boosting will create distortion, not bass.
• Low distortion: the lower the harmonic distortion, the more headroom you have for EQ adjustments.
• Planar magnetic drivers: planars generally take EQ better than dynamic drivers, especially in the low end. They tend to have flat, extended bass response with low distortion.

The Sine Sweep Test

Emrah's recommended test: play a sine wave sweep from 20 Hz to 150 Hz with a 10 dB boost applied. Listen for distortion or rattling. If the sine wave stays pure and clean, your headphone handles low-end EQ well. If you hear buzzing, rattling, or the tone breaking up, your headphone is hitting its limits.

Headphones That EQ Well

• HiBy (most models) and Ananda Nano — take EQ beautifully
• Audeze (most models) — handle EQ like a champ
• FiiO FT1 Pro — budget option (~$200), open planar, takes EQ very well
• FiiO FT1 Pro (closed-back variant) — also handles EQ well

Headphones to Avoid for Mixing

• Beyerdynamic (most models) — do not take low-end EQ well
• Sony MDR-7506 — no usable sub-bass or air; boosting produces nothing
• HiFiMAN Ananda (original, closed) — very difficult to EQ for mixing