Using EQ while mixing to compensate for frequency response of monitors/headphones

Question

I've been wondering for a while whether it is possible to compensate for uneven frequency response in a pair of monitors or headphones.

For example: I use the KRK Rokit 6's during my mixing as well as the AKG K701 reference headphones. If I knew the frequency response of my monitors/headphones, I figure I could correct the unevenness with an EQ on the master track, and mix with that EQ, only removing it before I bounce to audio for mastering. That way the response of my monitors/headphones would be flat during mixing (hypothetically) and I could get a better mix.

For example, here's the frequency response curve for the AKGs:

If I could use a compensatory EQ on the master that boosted the low end, lowered the low-mids a bit and lowered the 6k-9k range, would I not potentially be hearing a flat mix?

This seems like an obvious solution to the problem of coloured response but I have a sneaking suspicion it's not something that generally works. Can anybody explain to me what the potential shortcomings of this approach might be?

Answer 1

It is possible in some sense, and I believe there is software around that is designed specifically for this, but there are also reasons why this might not be the best approach. Obviously the character of your speakers affects the sound dramatically, but so to does the room. Measuring the response accurately enough to implement the EQ effectively would really take some doing, and even then, its likely to change change depending on your position. This kind of thing might be more effective for headphones, but if you're using them as an alternate reference maybe its best to maintain their spectral character?

I don't think I'd be alone in suggesting that the best thing to do is get to know your system and how it compares to others. Every setup will take some getting used to, even if you have really high end stuff with a comparatively flat response.

Answer 2

Systems such as SMAART and other Real Time Analyser based approaches can be used to level a system against EQ variation at a particular output level, however the problem is that the variations in a speaker can often be somewhat dynamic.

The exact behavior of a speaker is going to be a combination of how accurately it's voice coil is able to control the movement of the cone as well as the design of the cone and resonating spaces in the enclosure (among other factors). These will vary with both intensity of output and the presence of any constructive or destructive frequencies. In order to truly compensate 100%, you would need a full model of the response of the speaker to any combination of frequencies at any level and adjust your processing on the fly.

A speaker designed to be neutral has had choices made throughout the design to optimize it to produce as little coloring as possible. Trying to hack that in to a speaker that was designed to introduce coloring to the sound will be difficult because the you will be fighting the variability of the level of impact the design has under different conditions the entire time.

This doesn't mean it isn't worth doing if you don't have a better option available. Being tuned towards level for the particular level you most commonly are analysing at via pink or white noise playback would be beneficial, but it still won't be as good as a good set of truly neutral monitors.

Answer 3

As said in a comment above the measuring would be really hard if you do it by yourself. There is already software for this though: http://www.sonarworks.com/headphones

The software basically does exactly what you asked for. You have a separate measuring tool which generates a file that you can import in a vst plugin that you put on your master channel. As far as my comprehension goes the vst plugin only acts as an EQ.