[theheadfier]

This is my first thread here, and I figured I'll just share what the concept of "ideal" headphone really means to me.

Let me start by straight away saying that absolute tonal neutrality of playback equipment is at the very bottom of my priority list.

I noticed that a typical audio equipment review puts a lot of emphasis on "tonality", whether it is neutral, U-shaped or whatnot. On good reviews, this is done within the context of associated equipment as well as the specific music used. Here's the problem, though: If I'm reading a headphone review, for example, I don't have the same amp, DAC, and source used in the review. And more importantly, I don't listen to exactly the same playlist as the reviewer's. So with my own equipment and music, I might not get the same tonality from that headphone.

Ok, it's always better to actually try out the headphone with one's own music. But even so, a specific sound signature could seem perfectly "neutral" for some albums I have, but sound totally thin or boomy, too dark or too bright, on the other ones that I listen to. The disparity begins in the recording process. Many different factors at play here: microphone sound signature, capturing technique (closed mic, ambient, etc.), monitoring equipment, artist's creative choices`, and the sound engineer's judgment. All of those determine the resulting tonality of the recorded sound, and even so-called audiophile recordings don't sound the same. I also think that the "depends on the genre" thing is a silly generalization.

Instead of idealizing "tonal neutrality" when choosing headphones, I just pick the one that sounds convincingly natural with the majority of my music, and use EQ for the remaining few. Yes I do use EQ for those few other tracks in my collection, as I'm not one who buys a different headphone for a different sound signature.

So if tonal neutrality is at the bottom of my priority list, what is at the top? Well, I got a few things there. Things that really matter for good sound reproduction, regardless of how the recording sounds tonality-wise. Clarity and detail resolution, soundstage size, instrument separation/layering, dynamics, and speed/attack/decay… all these work together to provide great listening experience, and these things you just can't EQ your way into, and even the best DSP can't compensate for lack in any of these.

 

[bigshot]

You are correct that the type of music you listen to makes a huge difference in response. Balanced frequency response is essential for classical music and jazz, because they are largely performed by acoustic instruments that are recorded and mixed to sound natural. For rock music, response curves are all over the map. Electrified instruments are tweaked in all directions and mixes are often slapdash and poorly calibrated, particularly in current recordings made in home studios. 70s albums often sound muffled on the high end and current hip hop usually has bloated bass that makes it impossible to play on a system capable of reproducing frequencies below 50 Hz accurately. There's no one set of headphones that sounds good with all rock albums. The solution there is a set of headphones that has enough of a range that they can be EQed into sounding good correcting for the error in response in each particular album. I have a media server that has a section full of rock music, but I can't put it on random play because the engineering is so bad, I end up adjusting for every track that comes up. It's very frustrating. My jazz and classical libraries have music dating back 100 years and almost all of them sound right with a perfectly balanced response curve.

That said, neutrality is a great starting place. It's good to have a solid foundation to EQ against that you can always pop back to easily. You learn a lot about response when you EQ, so the more you do it, the better you will be at achieving the sound you are aiming at. In fact, you'll probably find that because of things like auditory masking, you'll end up shooting for a neutral balanced response naturally.

Actually, you might want to google up info on auditory masking, because response can have a huge impact on detail and separation as well. Response is the first priority to deal with, because when you can EQ that properly, it solves just about everything else except for dynamics. (Also headphones don't have soundstage. Speakers do. Headphones can't project at a distance of vertically. They project a straight line through the middle of your head.)

 

[Strangelove424]

So if tonal neutrality is at the bottom of my priority list, what is at the top? Well, I got a few things there. Things that really matter for good sound reproduction, regardless of how the recording sounds tonality-wise. Clarity and detail resolution, soundstage size, instrument separation/layering, dynamics, and speed/attack/decay… all these work together to provide great listening experience, and these things you just can't EQ your way into, and even the best DSP can't compensate for lack in any of these.

I understand the point you are making and agree with you to a certain extent. Frequency response can be EQed, and is much more flexible than most people treat it. I am a big believer in EQ and DSP, and adapt settings from album to album. On the other hand, most of the priorities you listed (detail, soundstage, layering, speed, attack, decay) can be interpreted through frequency response. They are different ways of describing the same thing.

Neutrality is not an ideal of its own (except for musical idealists I suppose) instead I think the purpose of getting neutral gear is to sound good with as many genres as possible. Typically a pair of cans that are colored for a certain kind of music (let's say Grados and rock) don't end up being good for other genres (like classical or hip hop) because their frequency is too far tilted. Yes, you could EQ them, but will eventually hit the limits of what the headphone was originally tuned to provide. I like to start from a neutral headphone and then provide my own coloration or DSP as I see fit.

 

[castleofargh]

the issue with neutrality is that it can be user specific and song/gear specific. so some universal signature labelled as neutral on headphones, well that's not very relevant. but at the same time, as the signature impacts pretty much about everything vaguely related to our impressions, it's hard IMO, to dismiss the FR on one hand, and list related stuff as more relevant.
and there is now little remaining doubt that your perceived neutral is the way to notice the most out of your music. so maybe you shouldn't be so fast to dismiss the signature. just work toward finding your neutral instead.
about different albums, musical genres etc, well that's indeed trouble. http://seanolive.blogspot.fr/2009/10/audios-circle-of-confusion.html

 

[theheadfier]

Thanks for the responses so far.

The good thing about great headphones is that many of them already have fairly good tonal balance to begin with. So that makes it easy for me to instead focus on the characteristics that matter more to me.

I beg to disagree that the criteria that I listed (clarity and detail resolution, soundstage size, instrument separation/layering, dynamics, and speed/attack/decay) can be addressed by FR and EQ.

A headphone that is naturally lacking in clarity/resolution due to high inherent distortion, for example, can't be corrected by EQ short of simply faking details (e.g. exaggerating the treble). Black background cannot be dialed-in using EQ.

Soundstage size (the "spread out", not a speaker-type soundstage which is impossible with headphones) can be faked by raising certain frequencies by EQ, but it's not the same as properly directing the sound e.g. driver angle (although this one is debatable).

Instrument separation/layering on a headphone with bad left-right driver matching, for example, is hopeless, no matter how you EQ that.

Speed/attack/decay depends on how fast the driver can start/stop. Again, no EQ can improve that.

I think there was no contention about dynamics.

As to shaping the FR for certain music using EQ, I certainly agree that there are limits. Therefore, I should have added to my list the "EQ-ability" of the headphone. In most cases, this is helped a lot by very low distortion levels across the frequency bands. I won't mention brands, but I had a few high end headphones that had gross amounts of distortion in the bass, and that's prior to any EQ.

 

[castleofargh]

EQ obviously isn't a panacea and should really be used to do what it does, EQ. on this I'm with you.

but again, subjective perception comes from the all sound package, so changing the frequency response can in fact affect many other things and at least in a few instances, help improve more than just the signature.
clarity/resolution, are terms I don't really get used that way so I'm not sure, but masking is a thing and it's fairly easy to make an area or a specific instrument better heard, by boosting the specific area of course, but also by lowering the frequencies just before it to reduce the masking effect and maybe get more details thanks to that. ultimately most serious work on this matter tends to suggest that perceived flat is how we can hear the most information out of a song. so FR is at least an element perceiving as much as we possibly can.

some aspects of localization do involve the frequency response, there is nothing to say that FR alone would be enough to fool the brain, but it still is an essential part of how we perceive placement in space.

I don't see why left/right mismatch a headphone couldn't be addressed with EQ. we just need to pay attention to the type of filter used to avoid pushing stuff off center, but that's about it.

speed/attack/decay, again if the amplitude the driver has to move changes due to EQ, in a way so does everything else. it's not a solution, EQ isn't the first thing I would look into if I wished to fixed damping issues for example, but we can't just put variables aside as if they were independent if they're not. and in a subjective point of view, perceived speed of a headphone is IMO more dependent on the FR than on actual speed. all the fast bass jokes I read all year long, often describe the perception from the FR and not some driver speed or how damped it is.

dynamic, I assume that once again it's used in a vague impression of dynamic, kind of way. because I'm not sure it makes sense to discuss the dynamic of a headphone aside from looking at how loud it can go. so if we're talking about how the headphone feels dynamic, then again FR is a massive element in creating that feeling.

FR is a big part of subjective impressions, and while changing the FR might only exceptionally be the fix for other issues, on the other hand, issues with the FR will undoubtedly affect how everything else is perceived.

 

[ev13wt]

I like a flat sound signature, in a headphone that resonds well to EQ. So a nice impedance curve and EQ.

Without a "baseline" flat - adjustments make no sense to me.


Same with speakers. I aim for flat, then EQ to taste-.

 

[theheadfier]

For loudspeakers I also like flat (as measured) FR as a baseline. For headphones, it's a lot more complicated than that, since a perfectly "flat" FR would sound like listening to speakers in an anechoic chamber, because there is no "room" for the sound to interact with. Nobody in the industry seem to have agreed on what is the ideal FR for headphones to mimic the effect of a flat measuring loudspeaker. Therefore, it's hard to tell from measurements which headphone FR graph approximates a loudspeaker's flat FR. Seems headphone reviewers use varying compensation curves in their published FR.

 

[bigshot]

Re: clarity

Again, try googling auditory masking. Clarity and definition of the separation of instruments are directly related to frequency response

 

[theheadfier]

I think I know what auditory masking you're referring to. I simply think that clarity can also be degraded by the presence of distortion among other things, frequency domain masking included. So really in terms of clarity, I'm more wary of the former (amount of distortion) which is an intrinsic quality of the headphone in question that can't be reversed by any DSP.

So really I don't disagree with you there.

 

[gardibolt]

My personal take is that I want everything up the chain to be as neutral as possible; then I pick the headphone to match the mood for how I want to listen. That way I eliminate variables and the fool's errand of trying to get the perfect match or synergy, if there is such a thing, and can concentrate on the one controllable variable of the headphones themselves. I tell myself that this saves me money, though I'm not sure that's possible in this hobby.

 

[bigshot]

I think I know what auditory masking you're referring to. I simply think that clarity can also be degraded by the presence of distortion

Yes it can, but almost any electronics you're apt to buy will have inaudible levels of distortion, and it really isn't much of an issue in good headphones either. Odds are if there is a lack of detail, it's due to masking problems caused by frequency response imbalances. The way to address this is to adjust the frequency an octave below where the masking occurs. That usually fixes it. Spikes in one frequency will mask the frequency above it. The midrange is the most common culprit, and you'll find boosted mids in certain brands of headphones.