How to equalize your headphones: A Tutorial

[Br777]

id say get an eq for your car. many cars can have them installed at a car audio shop..   get as many bands as possible so you have the most granular control

 

[ironmine]

Quote:

does anyone have any suggests on going about EQing othermedia? say example my cars speaker system

 
You need a notebook, a microphone and the program called "Room EQ Wizard" and a lot of patience and desire to learn
 
Do you have a parametric equalizer in your car?

 

[ironmine]

Quote:

this is the response that I currently have for my STAX headphones:

Mode: Linear Phase
Peak type: S-plane type II

(I always use the same mode and peak type, BTW)

Are you sure that it's better to use Linear Phase equalizing for fighting resonances? At least when I was doing DRC (digital room correction) using Room EQ Wizard, I was told by JohnM, who is the author of this program, the following:
 
Question:
"Which EQ mode should I use for room correction? E.g., eQuality VST EQ has the following modes: Digital, Linear Phase, Minimum Phase, Analog Phase."
 
Answer:
"Low frequency modal resonances are minimum phase phenomena, as are the parametric EQ filters which are best suited to addressing them."
 
(www.hometheatershack.com/forums/rew-forum/31428-questions-about-rew.html)
 
Another guy at GearSlutz.com also replied to me in the same way:
 
"I'd go for a minimum phase eq. With room EQ you're trying to address the minimum phase (invertable) aspects of your room and want to affect both phase and amplitude."
 
(www.gearslutz.com/board/studio-building-acoustics/515579-linear-minimum-phase-equalizing-drc.html)
 
Also, see this link: www.hometheatershack.com/roomeq/wizardhelpv5/help_en-GB/html/minimumphase.html#top, it explains why you can compensate with boosts for frequency response dips in minimum-phase regions only.
 
Headphones are like small rooms with similar resonating problems, so aren't we supposed to use the same techniques equalizing them? Especially if resonances are caused by modal problems. A room's fundamental resonant frequency can be calculated by dividing the speed of sound per second (1130 feet/sec or 344.65 m/sec) by twice the length. My room is 3.05 m long, so 344.65 / (2x3.05) = 56,5 Hz (microphone measurements confirm it). Using SineGen I found a peak centered around 7 kHz in my Denon D2000 headphones. Most headphone users complain about biggest peaks in 6 kHz - 8 kHz region. It means these resonances are created by the lengths 2,9 cm - 2,2 cm somewhere inside headphones. See the trend?
 
Try this: after finding the resonances (frequency, amplitude and Q) in your headphones using a procedure suggested by PiccoloNamek, set up a filter manually in Room EQ Wizard software, export it as a 32-bit impulse file (its better to export two versions: 44.1/32 and 48/32). Then use it in Voxengo Pristine Spaces (it's a VST convolver) to equalize your signal. Any other hi-quality convolver will do.
 
Tip: resample the obtained impulse files from 44.1/32 to 88.2 kHz/32 bits (for 44.1/16 music) or from 48/32 to 96 kHz/32 bits (for 96/24 music) using the highest quality upsampler available to you and feed to Voxengo Pristine Spaces the music signal upsampled to the same frequency as the impulse files.  Avoid non-integer upsampling! Use dither when needed.
 
This impulse convolving gives a more transparent result than using any VST EQ I tried (even my current favorite VST EQ - PSP Neon HR in FAT/non-LP mode).
 
Thanks PiccoloNamek for starting this thread and thanks to Lee Perry for directing me to it!
 
PS: If only we could place a small enough microphone in our ear (probably at the entrance to the ear channel or at the membrane?), we could use Room EQ Wizard for an accurate automatic measurement of dips and peaks in the frequence response of any phone!

 

[leeperry]

David Griesinger has a completely different technique: www.davidgriesinger.com/headphones.htm
 
It requires too much EQ to my taste, and I never good results out of it.
 
I don't think you wanna EQ the resonances in the cups(my cd1k cups are one big reverberating room), you just wanna kill the spikes you hear w/ SineGen(6400 and 9700Hz for me).

 

[ironmine]

Quote:

David Griesinger has a completely different technique: www.davidgriesinger.com/headphones.htm
 
It requires too much EQ to my taste, and I never good results out of it.
 
I don't think you wanna EQ the resonances in the cups(my cd1k cups are one big reverberating room), you just wanna kill the spikes you hear w/ SineGen (6400 and 9700Hz for me).

When I say "resonances" I mean "spikes".
 
I read this article: http://www.davidgriesinger.com/Binaural_Techniques.doc and it looks like things are more complicated than I imagined before, to say the least. E.g. (quote):
 
"TIMBRE AND FRONTAL LOCALIZATION WITH PINK NOISE
 
The apparent lateral position of the image can be easily adjusted by such a process, but it will probably appear to be either in the middle, above, or behind your head. It is much harder to make an image move to the front in the medial (up/down) plane. A pink noise signal through headphones usually sounds pretty bad in timbre, full of peaks and valleys in the sound One might think forward localization would im­prove if you removed these timbre errors with further equal­ization. Unfortunately it is not possible to simply adjust each frequency band until they all appear to be equally loud, since what you are really doing in this case is adjusting each band for equal masking. As Genuit has pointed out, masking de­pends strongly on the pinnae functions, and constant masking is not a good match to the frontal frequency response. In any case, if you think you know what pink noise should sound like when heard from the front you are probably fooling yourself."

 

[leeperry]

Oops yeah, I meant "middle ear resonances" not just "spikes"...anyway try both techniques and you'll see, but I prefer the OP for headphones EQ.

 

[sonance]

Quote:

Are you sure that it's better to use Linear Phase equalizing for fighting resonances? At least when I was doing DRC (digital room correction) using Room EQ Wizard, I was told ..

Hm, I was wondering the same thing, however I thought the peaks and troughs in headphones' FR were not necessarily caused by resonance, unlike in-room response compared to a speaker's known anechoic FR, which would often be caused be resonance and room modes?

 

[EYEdROP]

This thread has excellent info, so it deserves a bump. 
 
Tips for EQ noobs: 
 
1. Practice 
 
2. Patience 
 

 

[Elio]

I gave it a shot... was very skeptic to how much change would be,read alot... and did the EQ using sine wave, pink noise, Elecri-Q in foobar with a wrapper. 
heres what i got for my HD-515  : 

Used Linear mode but saw no difference between that and the analog mode. 
peak type I
 
1 : F : 3500Hz | -6 db | BW : 0.6 
2 : F : 7500Hz| -8 db | BW : 0.5
3 : F : 13870Hz | -7db | BW : 0.5 was surprised to hear that this Freq made a real difference in the sound. 
 
 
 
I have to say that the change is quite drastic...now the detail is much better and voices are much much clearer,and closer to realty especially in good recordings
i checked the new config with songs i know well form all kinds of genres and its made them all sound better. 
I'm happy with  this change. 

 

[leeperry]

the funny thing on the CD900ST is that it doesn't seem to require resonances EQ whatsoever.

 

[EYEdROP]

Quote:

the funny thing on the CD900ST is that it doesn't seem to require resonances EQ whatsoever.

Thats strange. Every headphone I have heard has atleast some peaks in the treble.

 

[EYEdROP]

Quote:

I gave it a shot... was very skeptic to how much change would be,read alot... and did the EQ using sine wave, pink noise, Elecri-Q in foobar with a wrapper. 
heres what i got for my HD-515  : 
 

I was skeptical too because of all the bad talk about using EQ. But with a parametric EQ, it makes it possible to zero in on problem frequencies only, without changing the other frequencies.
 

 

[EYEdROP]

Quote:

I was skeptical too because of all the bad talk about using EQ. But with a parametric EQ, it makes it possible to zero in on problem frequencies only. Its very precise. 
 

 

[nikongod]

Sennheiser said that the "dips" in the treble response on the HD650 was done specifically to combat some funky inner-ear resonances which is a big part of why these methods work so well. It would not surprise me that Sony tried something similar in the CD900st.
 
Quite a few people who have posted their EQ numbers take a few db out at around 5-7.5Khz, and notch in the ~10-15Khz range.
 
I would speculate that once you factor out non-linearities in FR of each headphone used I think things will average out to close enough for everyone that a generic solution is achievable which explains the Hd650 and probably CD900st. I wonder if we could find some common ground to test this idea on. I have thought of bringing my setup to a meet and trying this on a bunch of people to at least remove the headphone variable.
 
Something else that I think is interesting is how the EQ's that people are posting have the notches almost exactly 1 octave (double or half the frequency) apart.

 

[leeperry]

yes, the HD650 are said to be dark because they try to compensate for middle ear resonances.
 
on the cd3k drivers, I need to kill 2 spikes at 6400 and 9700Hz...and if you check this FR, it's almost spot-on at 6400Hz: http://www.geocities.jp/ryumatsuba/ath-a2000x.html

 
and the cd3k has a notch around 9500Hz: http://www.geocities.jp/ryumatsuba/mdr-cd3000.html