How to equalize your headphones: A Tutorial

[Joe Bloggs]

It sounds like you're used to the amount of treble energy in whatever cans you're using (which happen to be...? UERM customs?).  I would try adding a high shelf filter, say +3.5dB at 2500Hz, bandwidth 2 to bring the treble energy back up.  Basically the idea would be to have the curve stick to 0dB on average while evening out the spikes and dips in response.
 
Also have you played with the bandwidth setting on your filters, how are you arriving at your current settings?
 
Finally, the UERM are deep insertion customs right?  I wonder, with the depth of insertion of those things, perhaps the only real resonance you'd get is the one above 10kHz.  Then too, these being customs, I wonder if they'd have measured the distance to your eardrum while making the mold and taken that into account too with a custom filter in hardware

(doubtful, but I'd think it could be done) You might be the first guy to join this thread using customs, maybe you should shoot UE an email about this "resonance" thing and see what they say about it.

 

[Maxvla]

Yes I am EQing my UERM customs. I'm using .5 bandwidth which is recommended somewhere in this thread. Also surprised I'd be the first doing this with customs since I'm relatively late to the party with them as I just got them a few months ago.

As you can see from my adjustments to flat, these are pretty darn neutral already. I'm guessing the processing in this addon is not up to snuff because I do lose significant detail, and not just in the treble.

 

[Joe Bloggs]

You can't really settle on a standard bandwidth, it has to be customized to the actual bandwidth of your resonance.  Perhaps there is some causal relationship between the frequency of the resonance and the bandwidth, but if so I haven't seen it articulated in this thread.
 
Easiest way would be to get VAC+VSTHost working to filter the output from Sinegen in real time, but failing that, you can generate custom sine sweeps between resonance frequencies to determine the bandwidth required.
 
www.head-fi.org/t/587703/how-to-equalize-your-headphones-a-tutorial-part-2
 
details how-to.
 
(was going to write the updated guide you were talking about too...)
 
For the phones I've tried, the bandwidth of the resonance usually goes down at higher frequencies, as sharp as 0.2.
 
I'd be surprised if the quality of the plugin has anything to do with what you're hearing--much more likely an improper treble response is affecting imaging all across the board.
 
Then too I'm wondering how deep those customs go.  The minimum frequency at which it is possible to have ear canal resonance corresponds to a wavelength double the distance from your phone tips to the eardrum.  I would also try cancelling all the filters except the one above 10kHz and see how that goes.

 

[AstralStorm]

This is wrong, you can't just "remove resonances" - they're part of your head transfer function.
I'd skip the first post and just equalize pure tones or bandlimited noise bursts.
Sine sweep is actually less sensitive to my ears, I make mistakes when placing peaks using that.
 
Best done with foobar2000, since you can trivially adjust the frequency and reference signal level. (I use -6 dB and loudness or around 35 phon - of normal conversation.)
 
Your peaks look awfully wide, which is likely the same precision problem I had with sweeps.
 
 

 

[Joe Bloggs]

Why do we seem to disagree about every single thing in this thread?

There‘s a gentle, wideband resonance due to the ear canal (part of the HRTF), then there's sharp, narrowband resonances due to having your ear canal stopped up by an IEM and the resulting sound waves bouncing back amd forth between the IEM and the eardrum. I sure don't get the same peaks sing speakers that I do using IEMs.

He stated he used BW 0.5 for all the dips based on a recommendation. Sounds like he just found the peak frequencies and amplitudes then dialled them in reverse on Electri-q using BW 0.5. He needs a way to determine the actual bandwidth for himself.

 

[AstralStorm]

The resonances aren't just due to ear canal being "stopped" - there are open designs you know that don't have this problem, such as FutureSonics Ear Monitor which combines absorbing material with many vents. (Qts on the level of a small subwoofer!)
 
The resonances are usually simply the inverse of earlobe compensation and flaws in the driver.

 

[Joe Bloggs]

Ya know, for a second there I thought you were on to something scientific.  "Earlobe compensation"... then I googled "earlobe compensation" and found you'd performed a googlewhack--your post is the only hit!

 
OK, seriously.  Are you trying to say that the equal loudness curve for a loudspeaker placed in front of you is a smooth curve while the actual frequency response at the eardrum is a jagged response--that we psychoacoustically compensate for the occlusion effects of the earlobe and smooth out that frequency response in perception, such that when we wear IEMs we'd actually get a smooth FR at the eardrum but a jagged perceived response because the brain is compensating for the earlobe that acoustically doesn't exist anymore?
 
That almost sounds plausible... except what happens when sound reaches the ear from all directions?  The occlusion effect from every direction would be different and the notches in the FR (if notches they are) would be in different places.  Not knowing the source sound direction a priori how does the brain apply a general correction filter that fits sound coming from any direction?
 
Then too, if the above were true, playing sine sweeps with an IEM and evening out the peaks you hear would have the effect of restoring those missing notches the brain is trying to compensate and give you a frequency response congruent with sounds coming from outside the ear!  Wouldn't that be just great?

 
Speaking of absorbent, when I switch to comply foams on my SHE3580 I do hear two resonant peaks almost disappear.  (but comfort and uneven seal makes them hard to recommend for me)  But surely
 
1. you can't absorb all reflected waves--unless you block off the driver from the ear too.  The remaining resonances I hear are probably reflections from the driver itself.
2. whether the phone is open or closed has nothing to do with the issue.  The phone is open or closed *behind the driver*.  Sound waves are produced from the *front* of the driver, hit the eardrum, bounce back, and bounce back from the *front* of the driver.

 

[AstralStorm]

Quote:

1. you can't absorb all reflected waves--unless you block off the driver from the ear too.  The remaining resonances I hear are probably reflections from the driver itself.
2. whether the phone is open or closed has nothing to do with the issue.  The phone is open or closed *behind the driver*.  Sound waves are produced from the *front* of the driver, hit the eardrum, bounce back, and bounce back from the *front* of the driver.

 
The driver is almost acoustically transparent - typical membrane is far too thin and stiff to reflect or alter sound hitting it. (Calculate the wavelenth of the typical membrane which is les than 1/16 inch...) Perhaps it could cause resonance in 16 kHz range. BA drivers are even less likely to cause this.
 
However, the case is far thicker and will reflect in the higher frequencies both the sound reflected off the end of the ear canal and the "backblast" from the driver.
The interference pattern that forms causes peaks. If you make the other end non-reflective, you effectively remove the second part of the interference, resulting in a better sound.
 
Thus, an ultimate IEM would have both a seal between driver and ear and a completely open back - low isolation and low efficiency. Suprisingly, this wouldn't affect the bass response, should even improve it. ("infinite" Qts)
 
Quote:

Not knowing the source sound direction a priori how does the brain apply a general correction filter that fits sound coming from any direction?

 
Yes, that's one of the ways we detect the direction and position of the sound source at high frequencies. Low frequencies are unaffected by earlobe and pass through the head, so only interaural differences matter. It's part of the reason why intensity (point) stereo works so well - it doesn't really capture inter-aural delays (phase), but captures all the other components.
 
And there are advanced positioning systems that do take this into account for a truly realistic positioning. However, for music playback, you want a source that's placed directly in front of you. This happens to be a "flat" response, suprisingly. Add a low frequency crossfeed and you get a source very similar to one emulated by two speakers placed on the edges of an equilateral triangle and directed at your ears in "near field", so that there's no interference between them - point stereo.
 

 

[Joe Bloggs]

The driver is almost acoustically transparent - typical membrane is far too thin and stiff to reflect or alter sound hitting it. (Calculate the wavelenth of the typical membrane which is les than 1/16 inch...) Perhaps it could cause resonance in 16 kHz range. BA drivers are even less likely to cause this.

You kidding right? If that were true the driver wouldn't even produce any low frequencies when moving.

Thus, an ultimate IEM would have both a seal between driver and ear and a completely open back - low isolation and low efficiency. Suprisingly, this wouldn't affect the bass response, should even improve it. ("infinite" Qts)

Sure, since the driver seals the ear canal it‘s effectively an infinite baffle with the whole outside world as the enclosure.

And there are advanced positioning systems that do take this into account for a truly realistic positioning. However, for music playback, you want a source that's placed directly in front of you. This happens to be a "flat" response, suprisingly.

What do you mean by “flat“? It sure isn't psychoacoustically flat (equal loudness curves) and flat at the eardrum is even more implausible.

[/quote]

 

[AstralStorm]

Quote:

You kidding right? If that were true the driver wouldn't even produce any low frequencies when moving.
Sure, since the driver seals the ear canal it‘s effectively an infinite baffle with the whole outside world as the enclosure.
What do you mean by “flat“? It sure isn't psychoacoustically flat (equal loudness curves) and flat at the eardrum is even more implausible.

 
The driver can easily move air due to the seal - that doesn't mean it reflects sound well. A membrane is not a reflection dish.
And actually yes, psychoacoustically flat is an in-front central source, however certain frequencies are more important to get right - typically the closer to high sensitivity 3.5kHz region it is, the more important it is for localization. At highest freqs phase effects again start to predominate.
 
Edit: Someone wrote hell of an article on Wikipedia: https://en.wikipedia.org/wiki/Sound_localization

 

[Joe Bloggs]

Funny that sounds don't appear more out of my head after EQing out the spikes in the treble then.  If an in front source were to have a flat perceived FR...
 
Where are you coming up with this stuff anyway?  References please... and what's a technical term for "earlobe compensation" that won't result in a googlewhack?
 
How do you "Calculate the wavelength of the typical membrane which is less than 1/16 inch"?  Would these calculations happen to be much affected by the fact that the membrane happens to be a driver held in place by a suspension, with strong returning force?

 

[dustdevil]

Awesome tutorial, very detailed.
 
I thought my Sennheiser hd 555 sounded good before, and now its even better after lowering the 8khz about 5db.

 

[maxdot]

I'm sorry but I must be an idiot for not understanding the op "step by step" guide. I find it a bit abstract and complex to be frank. I've reread it several times and also read 10 first pages.
 
I have the sinegen program, and I got some pink noise in spotify. As I understood it, I in step 1 supposed to play the pink noise track and simultatiously press power in sinegen and drag the level up and down to find out which frequencies that sound louder than the others?
 
Btw the sinegen beebing sound is extremely loud for me default and I can not have my earphones on unless I lower it, so I just lower it to a normal volume right?
 
edit: So I ran the sinegen, just the sinegen no other sound (after setting it to loud but okay volume), and found 3 frequencies I find especially harsh. So I managed to do this in the Electri-Q program:

I just dipped down those exact frequencies a tad as you can see. I have no idea as to how shape that dip better than that, just pure luck really I found a way to make them that sharp actually after skimming thru the manual of the eq.
So, does this look correct?
I haven't used the pink noise yet though for the eq only the sinegen to find the freq.
I am able to have higher volume now after this. It does sound strange somehow though. A bit muffled the sound I must say...
edit2:
Tried to hear in sinegen how broad the harsh peaks were and came up with this:

(Realised doubleclick the dots u can decide deepness and depth of the curves etc by entering numbers doh. A bit easier
And this sounds much better than the last one. Still not sure if I'm on the right track here.
edit3: Not sure if I think it sounds better or worse hmm... It sounds a bit more muffled although at the same time more powerful since you can turn up volume more, and therefore you hear background sounds much better.

 

[Joe Bloggs]

I‘d say you're on the right track yes, can't help you much right now I‘m on a phone. Will see when I can talk some more about this...

 

[maxdot]

Quote:

I‘d say you're on the right track yes, can't help you much right now I‘m on a phone. Will see when I can talk some more about this...

Ok... I couldn't download the sine wave file aswell and I wouldn't know how to use it, or what to listen for, if I did. Basically I'm just doing the above and trying to ajust level and width when watching movies etc so it sounds ok. Any help would be appreciated indeed.