How far can EQ really go towards truly equalizing headphones?

After reading a lot of threads about using EQ, using it more myself, and comparing my own headphones to some others, I have been struggling more and more to tell the differences between cans. The better I EQ headphones, the closer they sound to each other; obvious, I know, but I am shocked by just how close. I was fooling around with a pair of KSC75's I borrowed from someone and, with some pretty hefty EQing, I was able to get a very solid sub-bass response (in the 20-40 Hz range) from them, even approaching the the power of my Beyerdynamic DT770 Pros (80 Ohms), which are already well known around here for their powerful bass response, to put it mildly! I do notice a little more distortion with the Koss set but I wonder how much of that would be reduced with even more precise EQing. And this has got me thinking.

My questions are: assuming that two headphones have been EQed to a perfectly identical and level frequency response, as tuned to my ears and the resonances therein, what will the sonic differences be other than soundstage and THD? Other than analyzing graphs, are there any cues in the sound that are easy to focus in on that will provide clues to these differences, especially ones I can pick up on before rigorous EQing? And, what are the implications of the first two answers when it comes to testing, purchasing, and owning headphones?
 
-Bizzlebin

It depends on how drastic an EQ is necessary. If you have to push transducers too far, they might distort or start messing up other frequencies. But I imagine most good headphones aren't that far apart.

Quote:

bigshot said:

It depends on how drastic an EQ is necessary. If you have to push transducers too far, they might distort or start messing up other frequencies. But I imagine most good headphones aren't that far apart.

Click to expand...

Sure. But if the said distortion is negligible―I get the ideal flat frequency response―what else besides THD and soundstage are going to be different between any given cans at that point?
 
-Bizzlebin

perhaps the time and phase of the driver itself? 

the decay of frequencies.
 
I love eq I really notice improvements with it with LCD-2 r2. One of the advantages is it's power handling and ability to handle bass eq. I noticed headphones like my denon d5000 and senn hd518 that they were  close or crossing the distortion threshold when reasonably loud.
 
Some might say micro detail retrieval. But that perception could have a lot to do with frequency response; highlighted treble.
Then you would also consider the ridgity of the driver material and how much as mentioned the decay is. I'm not sure you can always eq a sympathetic resonance out - if they even exist in headphones.
 
Benefit of LCD-2 they come with their own FR graph

Frequencies are what you actually hear, that's why frequency response is the most important factor in a transducer. As long as the driver can handle the EQ curve you're throwing at it without going splat, you should be able to make it sound however you want.

Distortion, dynamics and time based attributes will probably be pretty insignificant in most cases.

Quote:

udauda said:

 
FYI:
http://www.cpt.univ-mrs.fr/~briolle/11thAESpart1.pdf
http://www.cpt.univ-mrs.fr/~briolle/11thAESpart2.pdf
 
Indeed, we can certainly save some money.

Click to expand...

 
 
for more fun think about what the Smyth SVS Realizer is doing when it makes your headphones sound like the loudspeaker/room system you calibrated in - you can also calibrate, use different headphones with the same personal HRTF room calibrations
 
http://smyth-research.com/
 
what makes you think "the headphone's soundstage" can't be manipulated?
 
 
 
Quote:

jcx said:

"minimum phase" linear filters - like most analog EQ, and DSP IIR filters actually compensate for the minimum phase shift associated with amplitude variations in the transducer - EQ for flat frequency response and you improve the phase response
 
multi-driver transducers and the output above cone breakup at higher frequencies aren't "minimum phase" so linear EQ doesn't fully correct the "excess phase" part of the response
 
the "excess phase" response can in principle be fixed by more complex filters with all-pass phase EQ sections or in DSP FIR filters - usually you need automated measurement - our ears really aren't sensitive to phase above a few kHz
 
I do think the SVS Realizer is relevant to the discussion of what EQ can in principle do - even if you want to make distinctions between increasingly general "EQ"
 
1st order - a single frequency response adjustment curve - common to both R,L channels
 
next - custom per channel - even manufacturers bragging about matching their transducers only claim "1 dB" matching - clearly above ABX DBT detectable difference thresholds
 
cross-feed is simply "matrix EQ" - uses signals from both channels, cross-feed is the start of hrtf simulation, allows aproximating your hearing free field sounds with both ears
 
Dolby Headphone, some other SW adds simple idealized models of speaker separation, room reverb
 
the SVS system skips the modeling and measures the sound at your ears in the sweetspot of a real room and loudpseaker setup - including surround systems up to 7.1, adds angular response with the head tracking
and uses the in-ear mics to measure the headphones you're using, applying EQ for each channel - you can have separate headphone corrections in addition to multiple loudspeaker+room personal calibrations
 
 
for me all of the above are fundamentally related - we are multiplying the input channels by frequency/phase correcting filter transfer functions, possibly in in a full matrix and calculating the corrected 2-channel output to the headphones, only the head angle tracking of the SVS takes in any "new" information - and that just allows interpolation between transfer functions measured during the in room personal calibration

Click to expand...

Quote:

matt head 777 said:

the decay of frequencies.

Click to expand...

 
By "decay of frequencies", do you mean the impulse response? Isn't this correlated with (or better yet, just a different way to look at the same data as contained in) the frequency response?
 
Quote:

bigshot said:

Frequencies are what you actually hear, that's why frequency response is the most important factor in a transducer. As long as the driver can handle the EQ curve you're throwing at it without going splat, you should be able to make it sound however you want.
Distortion, dynamics and time based attributes will probably be pretty insignificant in most cases.

Click to expand...

 
I don't know if I would say the distortion is insignificant; on the graphs I'm viewing, it can range from over 10% THD+noise to well below .1%. But that's part of my third original question: what exactly are the implications of this seemingly enormous range?
 
Also, elaborate on "dynamics and time based attributes". Which attributes do you mean, how are they measured, what are their effects, and can those effects be easily mitigated electronically?
 
Quote:

Originally Posted by jcx /img/forum/go_quote.gif

what makes you think "the headphone's soundstage" can't be manipulated?
 
 
"minimum phase" linear filters - like most analog EQ, and DSP IIR filters actually compensate for the minimum phase shift associated with amplitude variations in the transducer - EQ for flat frequency response and you improve the phase response
 
multi-driver transducers and the output above cone breakup at higher frequencies aren't "minimum phase" so linear EQ doesn't fully correct the "excess phase" part of the response
 
the "excess phase" response can in principle be fixed by more complex filters with all-pass phase EQ sections or in DSP FIR filters - usually you need automated measurement - our ears really aren't sensitive to phase above a few kHz
 

Click to expand...

 
Ok, you got me there, jcx; even soundstage can be adjusted on a given set of cans using hardware or software, even free software. So that potentially narrows the list down even further: are there any things that can't be changed on headphones other than THD? What are the implications of this?
 
As for fixing phase response with EQ, that seemed to be the logical outgrowth of the thread http://www.head-fi.org/t/566163/headphones-are-iir-filters-graphs and the article it spawned on InnerFidelity. And if I understand you right, are you saying that single-driver headphones would be much easier to perform this ideal EQing with because they produce more "minimum phase" sound?
 
-Bizzlebin

bizzlebin said:

I don't know if I would say the distortion is insignificant; on the graphs I'm viewing, it can range from over 10% THD+noise to well below .1%. But that's part of my third original question: what exactly are the implications of this seemingly enormous range?

Also, elaborate on "dynamics and time based attributes". Which attributes do you mean, how are they measured, what are their effects, and can those effects be easily mitigated electronically?

Click to expand...

THD of over 10% is mighty high. Most reasonably good equipment has a distortion level that is low enough to not really be an issue. Dynamics and time based attributes are even less of an issue with modern electronics. The big bugaboo in getting good sound is always the frequency response. Imbalanced resonse can cause all kinds of problems... Listening fatigue from spikes, unnatural timbre from peaks and valleys, treble masking from midrange imbalances, etc. If you can tame the frequency response, there's a good chance everything else will fall into lne by themselves.

bizzlebin said:

Ok, you got me there, jcx; even soundstage can be adjusted on a given set of cans using hardware or software, even free software. So that potentially narrows the list down even further: are there any things that can't be changed on headphones other than THD? What are the implications of this?

Click to expand...

I may be mistaken, but there are some limits when it comes to external processing/filtering (of whatever type) and changing radiation patterns. For example, you can never make the KSC75 present a planar wave to your ears (like you'd get from an electrostat or ortho). Just like you can't create the S-LOGIC effect without manipulating the enclosure design of the headphones (but afaik, you *could* create the effect if you spent the time fabricating the enclosure (or just frankenstein'd an Ultrasone)).

That, coupled with the inability to solve very large resonance problems, are probably the biggest issues you have along with the individual characteristics of the driver (like THD and phase).

Everything else you should be able to affect with precise/clever enough filtering/processing, and enough time. If you allow physical modification to the headphone as well, then yes you can probably change everything (in the sense that, you could easily place whatever chosen driver into whatever enclosure you like, following whatever orientation you like (even exotics, like S-LOGIC and Auranomic)) except for the wave front presentation (in that, again, you can't make the KSC75 radiate in the same pattern as an O2 or HE-6, even if you get the FR to look identical or very close).

I'm just basing all of this on speakers - you can get "flat and clean" from all manner of speaker designs assuming you deal with acoustic issues and make use of equalization, but you can't fundamentally change their radiation patterns or alter the limitations of their chosen design paradigms - unless you can dramatically modify the enclosures or change transducers (at which point you're talking about different speakers imho). For example, the Sansui SF speakers will always radiate in a 360* pattern, and without changing their enclosures fairly dramatically, you can't do away with that. However, you could place them in a room with treatments designed to complement that radiation, and EQ designed to address both their own FR problems and the FR problems resulting in the room, and likely end up with a "flat and clean" image at the seated position. I think you can do the same thing with headphones, but I think there's still going to be some fundamental differences due to radiation/dispersion due to driver arrangement.

This all assumes power handling is a non-issue, and that your amplification and output equipment can survive the results of dramatic EQ curves that may see huge increases in overall demand without resulting in tons of distortion or damage. This probably isn't too dramatic of an assumption methinks.

It would be neat to try the SVS out with a few different pairs of headphones. :rolleyes:

yeah radiation pattern is another factor say an orthodynamic vs standard dynamic. LCD-2 vs AH-D5000 for example.
 

In simple terms, the more neutral a set of cans are, the more accepting they are of EQ, from my experience. Which seems irrelevant since you're usually wanting to use EQ to make them more neutral to begin with. However, like I've found with my HE-6s, I can easily imitate other planar headphones like LCD-2/3s and all of the other HiFiMan `phones. It's pretty cool. Just hit my preset and BAM, I have some LCD-3s. Also, notching down produces better results than up, but then again, quality neutral headphones are very flexible.

That's because it's easier to start at a baseline than to have to correct for an imbalance from the start. Headphones that are miles away from flat are an indication that the manufacturer is cutting corners in the design and really doesn't care about good sound.

bigshot said:

That's because it's easier to start at a baseline than to have to correct for an imbalance from the start. Headphones that are miles away from flat are an indication that the manufacturer is cutting corners in the design and really doesn't care about good sound.

Click to expand...

I don't think that's entirely fair - "perfect flat" does not mean "perfect fidelity" in headphones; a lot of well regarded sets do not measure (nor sound) perfectly flat, and it's not the result of any corner cutting, but instead a manufacturer having a specific view/opinion/design schema that targets certain goals. An example is the Grado GS-1000 - which measures closer to the equal loudness contour than anything else. Is it the result of corner-cutting?

You also have to remember the ideal compensated response curve(s) for headphones do(es) not look flat, and there are many headphones that target them (e.g. K701).

Flat frequency response is better sound. Deviating from a natural EQ is what makes headphones tiresome to listen to for a long time, and can lead to masking issues which obliterate certain frequency bands. Sound engineers mix to a flat response. If you want what they intended you should listen to a flat response. A slight gradal slope is not as much of a problem, but it's still inaccurate.