Five EQ filters ought to be enough for any good headphone (or IEM?)

[IAtaman]

Is my current theory that I would like to bounce off of you gentlemen.

I suspect it might be obvious to many of you who are more knowledgeable and experienced than I am, but to me it came as a slow revelation as I spend more time trying to solve EQ issues using MacOS. And in full, it goes something like this:

It should be possible to tune a decent headphone to your liking using 5 EQ filters or less. If the headphone requires more than 5 filters to sound good, it is not a good headphone. And fine tuning with more than 5 filters is unnecessary nitpicking.

I have thought about this in connection with headphones, but maybe it might be true for IEMs as well?

Why five, well because RME ADI-2, but also I figured, to get satisfying results I usually need a sub-bass shelf, a bass correction (hate the 200hz boost), 2 mid/treble corrections and in some instances a fix for ear piercing highs.

In any case, what do you think?

 

[Blumlein 88]

Probably true if they are parametric, and you can move the frequency centers around and adjust the Q.

 

[IAtaman]

Probably true if they are parametric, and you can move the frequency centers around and adjust the Q.

Yes, true, I was thinking parametric EQ indeed that allows you to change freq and bandwidth or Q along with gain of course, and can implement high or low shelf.

 

[isostasy]

Amir's reviews probably fit this line of thinking off the top of my head. I can't recall any headphones he recommends that require more than 5 filters at least.

 

[Jimbob54]

You are probably right but there is a tendency I see a lot on here to want to EQ "to" rather than "closer to" target. Hence you see lots of very precise adjustments recommended by other members. This seems to be accompanied by a belief that all changes to an EQ profile to align better visually to target result in an audible improvement. This is not Amir's approach obviously

 

[solderdude]

Sometimes only 1 or 2 filters is all that would be needed and 5 would be overkill.

EQ'ing exactly to certain 'measurement' made on 1 specific fixture where it is unclear how the final trace was created is often seen as the holy grail.
An owner of a headphone will, with 100% certainty, have a different HRTF, seal and positioning than what was used to create that specific plot so the trace you see and base the EQ on will certainly deviate from what reaches your eardrum.

People should realize that measuring a headphone is NOT an accurate scientific endeavor and also not 'matured science' so any EQ based on it also is not accurate.
That EQ is only accurate to that specific 'wiggly' and can make that wiggly seem 'perfectly compliant' with their favorite target.

I am sure the vast majority of 'exact EQ' believing people believe it is more accurate though the more 'wiggles' are 'corrected'.
The reasoning in believing in such EQ is that 'it has been done by this or that person' or on 'this or that specific fixture' and referenced to [insert your favorite]curve so it MUST be accurate.

Of course, when one fiddles with all the 'filters' you will hear differences. That does not mean the headphone is actually MORE accurate on one's head it just means you can hear the differences and when one believes that is 'as accurate' as one can get will believe that is how that headphone should be.
It will, most likely' be closer to 'accurate' than when no EQ is used.

So... for this reason I think Amir's method is valid enough between 100Hz and 6kHz.

In fact, probably the most 'valid' EQ would be based on looking at various measurements of the same headphone taken on different fixtures and taking targets in mind and try to figure out the 'biggest' and wide bandwidth deviations found and create a simple EQ on that and subsequently adjust to taste (or seal, or positioning) from there.
That is a more scientific approach than believing in (banking on) 1 trace obtained in a certain way with a fixture they believe in and target they prefer made by one individual.

 

[markanini]

@solderdude suggested broad adjustments and I'm with him on this. A set of parametric lo-shelf, peak and hi-shelf filters to should be enough to account for individual fit and preference for bass levels. There's a sweet spot between adjustability and ability to engage the ears while doing so.

 

[kemmler3D]

And fine tuning with more than 5 filters is unnecessary nitpicking.

Hey man, don't you know where you are? This entire hobby is unnecessary nitpicking!

But in all seriousness, I don't agree with "5 filters ought to be enough".

If the headphone requires more than 5 filters to sound good, it is not a good headphone

This might be a decent rule of thumb, but you're not going to find universal agreement about the words "requires" or "good". If you want to get really literal - require means "need" and there's no true need for good headphones in the first place. Nobody will die without them.

Trying to draw a line in the sand where enough effort or enough quality is enough - when we're talking about genuine audible differences - flies in the face of the entire concept of hi-fi.

Other reasons I don't agree:

Oratory 1990 often suggests headphone correction EQ curves with 10 or more filters, I'm not going to hold myself out as more knowledgeable than them.

Even mid-tier Bluetooth modules come with DSP that enables 10-20 PEQ filters.

More EQ bands don't cost money.

More EQ bands aren't inherently harmful to the sound.

5 bands isn't nearly enough to correct more than broad defects... if you have peaks in the response, you're going to need at least one more band each. It's not uncommon to have 2-4 unwanted peaks in the response of any given headphone.

Case in point, I added 3 or 4 more bands to the EQ for my Focal Elex the other night, by ear. I'm not tuning to a target curve, but eliminating defects that I think I hear. Now, maybe this is just placebo, maybe I actually made things worse.. but even if so, I spent $0 and I like my headphones more than I did before. Isn't that the dream?

I would say enough EQ is enough when you can't hear anything in the response you want to change, or you're adding a filter and you're not sure if it's an improvement or not. That might be at 3, 5, or 15 filters depending on the user and the headphones.

Where I agree:

I would generally agree that if something still sounds noticeably bad after 5 noticeably helpful EQ filters, it's probably just bad.

 

[mmmdc]

You are probably right but there is a tendency I see a lot on here to want to EQ "to" rather than "closer to" target. Hence you see lots of very precise adjustments recommended by other members. This seems to be accompanied by a belief that all changes to an EQ profile to align better visually to target result in an audible improvement. This is not Amir's approach obviously

Differences across units and personal fit alone make that kind of a silly endeavour.

Unless you have your own measurement rig at home of course.