DJ
DJ Toolkit
Calculators and visualisers for producers and DJs
 Home
EQ and Monitoring

Headphone EQ Compensation

Every headphone has a voice, and the part of it that shapes your mix is the part you cannot hear until it is on a club system. A pair that lifts the low end and hollows out the upper midrange will make a mix sound finished on your desk and thin on a dancefloor. The fix is not to mix quieter or louder, it is to put the missing part back and take away the part that was never missing. This tool takes a measured response curve for the headphone you own, compares it against a target curve, and hands you five parametric EQ bands with the frequency, gain, Q and filter type already filled in. It also shows you the residual error afterwards, because five bands cannot fix eight independent problems, and it will not pretend that headphone compensation repairs a room.

Measured response of the model, reduced to eight bands
What the headphone should be doing instead
dB
Added on top of the correction below 120 Hz
dB
Added on top of the correction around 3 to 6 kHz
Result

How to use it

STEP 1

How to use this tool: pick your headphone model first, because the correction is meaningless without knowing which response you are correcting. Then choose a target. The Harman curve is the safest choice if you have no strong opinion, since it is the closest thing to a broadly agreed reference for what most people prefer. The flat reference is for engineers who want to hear the mix as it is, and the bass boost target is for checking a club track on the system it will actually play on.

STEP 2

Read the correction table before you touch a plugin. It shows the measured level of the headphone, the target level, the difference you need to apply, what the five bands will actually deliver, and the residual error that is left over. That last column is the important one, because it tells you which part of the headphone will still colour your mix after the correction and where you need to trust your judgement rather than the numbers.

STEP 3

Enter the five bands into any parametric EQ: frequency, type, gain and Q are given for each one. Start with the low shelf and the high shelf first and check the balance, then add the narrow bands. After entering them, play a reference track you know well and compare the corrected version against the uncorrected one, because the point of the exercise is not to reach a number but to make the headphone boring again.

STEP 4

Use the extra bass and presence fields sparingly. They add on top of the computed correction rather than replacing it, so a large value can push the low end or the presence band somewhere you did not intend. A value of one or two decibels is a useful nudge. Anything above four is a different target curve, and you are better off choosing the bass boost target and editing from there.

Three common mistakes

  • Common mistake: treating headphone compensation as a cure for a bad room. It is not. It removes the colour of the headphone and nothing else, while early reflections, modal ringing between the listening position and the speakers, and the way your monitors are placed stay exactly as they were. The common failure is a mix that translates perfectly on the headphone and falls apart on the system, which is a room problem that this tool cannot see.
  • Watch out for stacking a compensated headphone on top of a plugin that already has a sound, such as a gentle mastering EQ or a headphone plugin from the manufacturer. You will be correcting a curve that has already been corrected once, and the two curves rarely line up. If the headphone plugin ships with the hardware, use either it or this correction, not both, and re-check the level after switching.
  • Key point that gets missed: the residual error is where the mix will still be biased. Five bands cannot chase eight independent differences at once, so the shelves that fix the low end and the air also colour the bands next to them. Reading the residual column tells you which direction the rest of your session will lean, and that is often more useful than the headline correction, especially on headphones with a strong bass lift.
  • Easy to get wrong: correcting the headphone and then trusting the result completely. A compensated headphone is a much better monitoring tool, but it is still a pair of speakers in a room, in a position you chose, at a volume you chose. Cross check on a different system, check the mono fold down, and check the top end on a phone. The compensation gets you closer, it does not close the argument.

What the target curves actually are

The Harman free field curve from 2018 is the result of blind preference testing across many listeners and it is the closest thing the industry has to a shared reference. It carries roughly four to five decibels of low end lift around forty hertz, a small dip in the upper midrange around three kilohertz and a gentle rise in the treble. The diffuse field version has more bass and a flatter middle, because a phone speaker or a laptop reproduces the low end so badly that a strong bass shelf is what listeners reach for.

  • Harman free field is the balanced default for making decisions about balance and translation.
  • Diffuse field has more low end lift and is useful for checking a club track against a phone.
  • Bass boost is deliberately heavy and is a translation check rather than a mixing target.
  • Reference flat removes all tilt and is for engineers who want the mix, not a preference curve, though it is not what most listeners prefer.

Why five bands and not eight

Every correction here is computed on eight bands, but the output is five EQ bands, because five is what a usable parametric EQ and a fast session actually allow. The tool picks the five bands with the largest deviation, which means the loudest errors get fixed first and the small ones are left alone. The cost is that a shelf is never a single band. A low shelf at twenty five hertz bleeds upward, and a high shelf at twelve kilohertz bleeds downward, so fixing the extremes slightly changes the middle.

  • The residual column shows what survives after the five bands are applied, band by band.
  • A residual of one or two decibels in the upper midrange is normal and usually harmless.
  • A large residual means the headphone and the target disagree in a way that needs more bands, or a different target.
  • Adding a sixth and seventh band is fine if your EQ allows it, but each extra band makes the curve harder to hear as a whole.

Compensation, mastering chains and the limits of the exercise

Headphone correction is a monitoring tool, and it behaves like one. It gives you a flat starting point so that when you boost the bass on a track you are boosting the bass that is actually there, not the bass the headphone was already adding. It does not give you a translation guarantee, because translation depends on the playback system, the room, the format and the mastering decisions that happen after you export the mix.

  • Apply the correction to a monitoring chain, not to the mix bus, unless you are deliberately rebalancing the track itself.
  • Re-check levels after correcting, because the headphone output level often shifts with the EQ in place.
  • Keep the correction in a separate plugin so it can be bypassed for a final sanity check.
  • Pair it with a room measurement if the room is bad, because the two problems are independent and both have to be addressed separately.

FAQ

Is this measurement data accurate enough to mix on?
It is accurate enough to reason about and to build a starting correction, and that is all it claims to be. The curves here are eight band summaries of published measurements for each model, which means they capture the tilt and the broad emphasis but not the fine ripple between the bands. Two pairs of the same model also differ by a decibel or two from unit to unit because of the fit and the seal. Treat the numbers as a direction to move in, then verify by ear against a reference track you know.
Should I correct a headphone that already ships with a correction plugin?
No, pick one or the other. Manufacturer plugins are usually tuned for the intended use of the headphone, often for a flat or a slightly warm result, and stacking a second correction on top applies two different curves to the same headphone. The result is often a strange over-egged midrange that sounds analytical rather than accurate. If you want the manufacturer behaviour, use it. If you want to mix to a target, bypass it and use this correction instead.
Why does the residual error not go to zero?
Because five smooth filters cannot reproduce eight independent differences. A shelf that adds two decibels at sixty hertz also adds something at one hundred and twenty hertz, and a high shelf that restores the air at twelve kilohertz inevitably touches six kilohertz on the way. The residual column shows exactly how much of each error survives. If the remaining error is concentrated in one band, add a narrow peaking filter to take care of it. If it is spread everywhere, the headphone and the target simply disagree too much for five bands.
Does headphone compensation let me skip checking on speakers?
No, and the assumption that it does is the most expensive mistake in this area. A corrected headphone is a good translation tool for balance, for arrangement decisions and for spotting problems in the low end and the midrange. It is still a pair of small drivers close to your head, in a room with its own reflections, at a near field distance that no playback system will reproduce. Keep a check on a second system in the rotation, always check the mono fold down, and check the top end on a phone before you release.
Which target should I use for a club track?
Harman for the mixing decisions, because it is the most neutral option that is still broadly preferred. The low end will sound slightly lighter than you are used to on headphones with a big bass lift, which is the point, since most club systems will not reproduce that much sub anyway. Once the mix is done, switch the target to the bass boost curve and listen for translation problems, because that is roughly how the track will behave on a phone or a system with a small woofer. If the mix collapses there, the low end balance needs work.

Limits of this tool

  • Limits: eight band summaries of published measurements, not per unit measurements. Good for tilt and emphasis, not for fine ripple between the bands.
  • Not for: room correction, monitoring speaker alignment, or predicting how a mixing console or a plugin will colour the signal after the correction.
  • When this stops: models with strongly different shapes between bands, or headphones with a narrow peak in the upper midrange, need more bands than five to bring under control.
  • Scope: the correction is meant for a monitoring chain. Applying it to the mix bus rebalances the track rather than the monitoring, and it is a different decision with a different result.