Neutral Sound: Practical and philosophical Q

[earmonger]

Amps and DACs can provide measureably imperceptible distortion, passing trough the original signal.

Why not headphones and IEMs? I don't want fun, warm, dark, treblehead, basshead, chucklehead, etc.

I just want to hear what the musicians and producers put into the recording in first place. Who just does that? if not, why not?

 

[GREQ]

Well then... let's open this can of worms.

I think the short answer is that this 'journey' to find 'the one' headphone will never end for humanity.
The simple reality of things is that every human ear is different, and none of us can fully agree on what truly is 'neutral sound'.

I think this is why the community is drawn to certain headphones for certain genres.
Take the AKG K501 for example.
It has been highly regarded as a master of classical music because of it's soundstage properties combined with it's as-close-to-neutral midrange tuning as possible.
It does not have neutral bass/sub-bass.
BUT, because of that combination of bass roll-off, tuning and soundstage, it wonderfully mimics the feeling of being present in a concert hall.
It also does a great job with any acoustic and rock music.
Philosophically, this older headphone reached VERY close to the traditional concept of 'neutral'.

Let's go to the opposite end of the music spectrum.
JVC HA-SZ2000 is the only headphone ever made with a properly implemented 'sub-woofer'.
Out of the box, it sounds terrible.
You can't use this headphone without applying an insane EQ curve filter.
But what it can do with sub-bass is utterly unique - no other headphone can get anywhere close to it's ability to give you that oversized sub-woofer-in-a-car feeling.
It sounds like you're in a carpeted closet, but you're sitting on a giant cinema sub.
Philosophically, as stated by the designers of this headphone, it reaches to re-create the feeling of live music, specifically modern music with emphasis on bass.
Is that traditionally neutral? Absolutely not.
But is it able to more closely replicate a live club sound (specifically with live electro-recordings)? Absolutely yes - which is in of itself, philosophically speaking, closer to the 'neutral' - or the originally intended sound at a live club venue.

Neither of these headphones are 'neutral', but their ability to translate music into an experience that closely mimics previous and/or other human experiences are remarkable.


Next is how sound is created.

Headphones and IEMs exist (in various stages of development or times in history) with many different transducers:
• dynamic
• planar/orthodynamic
• electrostatic
• electret
• piezoelectric
• ribbon
• AMT (air motion transformer)
• BA (balanced armature)
• wing
• SDM (stretched duralmin membrane)

Each technology is aimed at resolving exactly the problem your question aims to find an answer to, but none of them are able to give that answer.
Every transducer type has some incredible 'pros' while being riddled with 'cons'.

Each will present sound in a different way due to their thickness, materials, shape etc - every variable counts.
Dynamics are usually small, light and conical in shape. Therefore they will present a different 'physical' sound wave than a planar or electrostatic, which are often much larger surface areas and are perfectly flat - so they produce something more like a 'wall of sound' which is only as diffused as the maget array or stators in front of it.

Dynamics often have 'tuning lenses' which tune and diffuse the sound wave further.
Many people prefer the sound of drivers that are 'naked', with as little diffusion or front-end tuning as possible.
It's fair to say that a unfiltered sound in this case can sound more natural, but then you have other modern headphones like DCA Expanse which use it's proprietary AMTS filter to VERY specifically tune the sound closer to a 'measurable' neutral.

Many drivers have also been 'doped' with other materials by a process of vapour deposition and coated with diamond, ruby, sapphire, carbon, etc.

The same goes for ear pad shape:
• over ear
• on ear
• inner ear (IEMs)
• resting in the concha (flathead ear buds)

And material:
• leather
• velour
• pleather (protein/artificial leather)
• alcantara
• memory foam
• (too many foams)
• silicone
• fabric mesh
• plastic

How will people ever agree on what drivers and what material sounds the most natural or neutral when there are SO many variables to consider?
The same goes for loudspeakers.
We have so many speaker and sub-woofer designs with differing driver counts, firing directions, giant horns, inner chambers, ports, etc ad nauseam... it won't end.

My Personal Conclusions...

I've owned and heard a few headphones over the years, and after some time I think it really comes down to two things.
• Personal taste.
• Your ears.
If you think a headphone sounds neutral, based on the music you listen to, the recordings you've heard, the live performances you've experienced, then it is a good headphone for you.
And if you like a headphone for it's ability to transport you to feeling or experiencing music in a unique way that you enjoy, then that is also a good headphone for you.

There are no wrong answers for this.
Unless it's Beats... that s***'ll get you in trouble around here....

Practical Conclusion:

The best transducer would have the following properties:
• Weightless
• Perfectly rigid
• 'Massive' excursion
• Zero distortion at any frequency at any given moment no matter how complex the music
... unlike amps and DACs, drivers have to conquer the resistive physical properties of air, the mechanical resistive properties of the materials they are made out of, and be scientifically 'tuned' to be the correct distance from, shape, size and ergonomics for all human heads.

It does not, and likely will never exist...(?)

What we really need is science fiction!!
Brain implants that bypasses our ear dums and directly feeds information from a DAC to our brain that is translated into audio.

 

[ProtegeManiac]

Amps and DACs can provide measureably imperceptible distortion, passing trough the original signal.

Why not headphones and IEMs? I don't want fun, warm, dark, treblehead, basshead, chucklehead, etc.

I just want to hear what the musicians and producers put into the recording in first place. Who just does that? if not, why not?

Because it's not absolutely intentional. They literally have not made a driver that has an absolutely flat response the way upstream electronics can becaus physics.

Let's look at a dynamic driver. Larger driver means it can produce more low frequencies at higher output levels, but as it pumps back and forth, that wider diaphragm flexes more, and when it needs to reproduce a higher frequency, it needs to pump back and forth faster, enhancing that flex on the diaphragm. That's why even if you EQ a signal so you can have very high freqs off a subwoofer but the response is distorted. On the opposite end of the spectrum, you have a tweeter that is tiny so when it pumps back and forth, it can do so at a very high rate and reproduce high frequencies without much abnormal flexing on the diaphragm, but it can't produce low frequencies since it lacks surface area. Then you have midwoofers, midbass, and midrange drivers taking up their own ranges in between these two. Headphone drivers are fullrange drivers that can work well enough ie wide enough range...if it's about a couple inches max from your eardrums. See, the response isn't even just where the mic measures it, it's also in how far you measure it at (you can test this by setting the headphone on a table and compare the sound to when you wear it), so even with speakers that can have practically all the size restrictions removed so you can have the biggest horns and the most intricate vent design to extend the response without muddying it up will still not even just not be completely flat, it also would not sound the same from 1m (as per usual measurements) as it would from well over 2m (I mean, if you have HUGE speakers, chances are your room is also huge).

Other driver designs aren't gonna drastically change that. Sure a planar doesn't have to pump back and forth hard, it just vibrates...and then your common HiFiMan tends to have a flat response from 20hz to 1000hz, nosedive, rise to a higher point at 3500hz, dive a bit, rise, dive, rise really high...in short that response above 1000hz wasn't HiFiMan intentionally giving listeners a weird response, it's a limitation on the driver design. When others intentionally tweak the design to affect the sound it's more a matter of compromise. Can't reach down to 20hz? It's OK, not like a headphone at the ears will have bass frequencies hitting the listener's entire body to enhance perception of the bass (ie why some people feel like something that hits more freqs evenly "lacks bass"), let's follow the Harman curve and deal with perception!" So they go boost the 65hz to 120hz region or therabouts and you end up hearing a louder thud from the drums that produce sound in that region along with some of the bass guitar or bass viola.

In other words...it's kind of like asking why there isn't a four door sedan that has the ground clearance of a sporty crossover and handling dynamics and acceleration of a low slung supercar. You can't have an M5, FXX, and Urus in the same car. It will always be a compromise and electric engines won't change that, more so if we're still on tyres; maybe with anti-grav drives you can have something the size of a minivan going as fast as current cars or something. Same thing with audio - current tech just isn't there yet.

 

[castleofargh]

Amps and DACs can provide measureably imperceptible distortion, passing trough the original signal.

Why not headphones and IEMs? I don't want fun, warm, dark, treblehead, basshead, chucklehead, etc.

I just want to hear what the musicians and producers put into the recording in first place. Who just does that? if not, why not?

Some issues we are facing:
-we listen with headphones/IEMs when albums are almost universally mixed and mastered on studio monitors. That causes several problems with us receiving stereo in a very different way.
-your own body is part of the equation, making it impossible to get one standard as electrically or acoustically flat/accurate. This again creates a bunch of extra variables that are a function of your very own body.
Speaking about frequency response, so it doesn't take all day, when a sound reaches your eardrum in everyday life (sans headphones), you have affected the FR a great deal with your very custom-made body. Just your shoulders can be enough to change sound in a way that your brain can recognize. The shape of your head does too and of course your outer ears, which tends to be fairly unique for a given person. Just put your hands behind your ears to get an exaggerated example of how much some flesh can reflect, direct and at some frequency, resonate with sound.
But that's just half the problem, because of course we're used to how things sound altered by our body, that's what we think of as normal hearing. Enters the headphone. Sound is now emitted straight at the ear, so it doesn't have the characteristic changes from our head and torso. And even the outer ear is giving confusing cues that are the same no matter the panning of the instruments on the record.
Because we're so used to hearing with sound altered by our body (check out HRTF) which we think is how neutral sounds, that when the headphone bypasses some of that, it feels like an EQ going the opposite way of the EQ normally applied by the body. It's extremely unlikely for a headphone to match the FR of even one direction by chance and random try of pairs.


For measurement’s sake, some people took measurements on a bunch of cadavers right at the eardrum, averaged that and made a coupler and dummy head supposed to come somewhat close to THE AVERAGE made up human head over a fairly restrained frequency range (most used one in history was given as human average up to 10kHz). It was nobody's exact head, but it was a usable repeatable reference which is incredibly useful for measurement. Then someone decided to create a competing product, then some coupler got made that was a little better, then one got made to be somewhat reliable above 10kHz. And soon enough we had too many competing standards. Of course those suffer from the same issues we do, they all give somewhat different results because they all have a somewhat different HRTF, coupler impedance, resonance, all the same troubles we have with real humans being different.
Even with one unique standard, there was no answer for what's neutral for you specifically, but it sure didn't help to have many standards.

You can try to work things out, but sadly basic EQ will only help for one direction at a time, and that EQ will benefit you and not me as I'll need another one fit for my body's acoustic impact. And other issues related to spatial localization often come in the way of proper perception, no matter how good the EQ. An example that hopefully makes a complicated matter more explicit: I mostly feel like the so-called neutral headphones put the singer on my forehead or even on top of my head. With EQ I should objectively be able to get the singer back at eye level as in terms of sound, it's just a matter of getting the FR that matches sound hitting the outer ear from a source at eye level. It's mono sound, so there is no issue with interaural time delay, FR should be what convinces my brain that a sound comes from a particular elevation angle. Sadly, it doesn't work for me. I can only guess as to why exactly, I suppose it has to do in part with how I'm not seeing the singer in front of me and my brain trusts my eyes way more than my ears. It can also be that anytime I move my head, the voice turns with me, and my brain interprets that as the singer being stuck to me. As I don't see him, he must be on top of my head where my eyes can't see him! I guess it could be something like that going on in my mind.
I solved that with the Realiser A16, but it's a mind-boggling overly complicated and overly expensive piece of equipment that really just does what speakers do better. I went for it because where I live, speakers would soon get me killed by mad neighbor grandmas with their hysterical hand sized dogs.
But you still need to make measurement using real speakers at least once (probably more because it's a PITA to do it correctly in one go), and even then, with what might be for now the best speaker simulator on headphones, you still should do some EQing to account for possible issues and ear canal resonance as the mic obviously isn't placed at your eardrum (that would hurt!!!).

As others have mentioned, headphones aren't remotely as good sound reproduction tools as DACs or amps. They are up to a magnitude worse for several variables. The fidelity of a given headphone is not great, and I don't think there exists a headphone that doesn't have some audible "character"/distortions(THD graphs are only about harmonic extras, which is not all there is). Some headphones might not even be good enough to allow for the EQ you need. A rule of thumb is to get a headphone with low distortion and a FR that's pretty smooth and well extended FR. Smooth anything can be EQed easily, spikes and roll off due to physical limitations, not so much.





There you have some of it.
-1/ If you want a sound close to what was intended, get speakers and get a guy whose job is to set them up and mitigate the various room acoustic issues that you have in that room.
-2/ If you want to stick to headphones, and you don't know what to do of your money, I'd say get an A16 and an objectively pretty good headphone, then join the A16 thread to get all the help you'll definitely need for speaker measurements. The nice thing is the possibility to record a bunch of rooms and speakers, and also to have multichannel in almost all formats (if you record enough speakers in all the necessary positions).
-3/ If you don't want to spend a stupid amount of money for a speaker simulator, You can invest in the cheapest binaural microphones (the stuff you put in your ears), download Impulcifer https://github.com/jaakkopasanen/Impulcifer and join that thread to ask for all the help you'll definitely need https://www.head-fi.org/threads/recording-impulse-responses-for-speaker-virtualization.890719/
It doesn't have all the decoding of movie standards or head tracking, but it's a free software, and it lets you know if this is the path you didn't know you were always wishing for. A low-cost opportunity to find out and maybe get blown away.
-4/ Or if you happen to like the "cleanliness" of headphone sound more than speaker sound, You can try some more basic crossfeed solutions and EQ (also needs to be pretty custom or a lot of luck as the settings also depend on your own head). This is kind of a midway solution to get sound that could feel closer to what was intended, but isn't that close to being it.
-5/ If none of that feel right even after trying a bunch of options with a bunch of settings, then I'd circle back to simply using EQ to get what feels right/flat for you. Starting with the Harman target for headphone has a little over 60% chance of saving you some time tuning the EQ, but most likely you'll still be able to do a little better with extra fine-tuning (like amount of bass and maybe treble depending on your age/ear damage and general HRTF). That's nowhere near the sound intended by the sound engineers listening on their studio monitors, but most people are satisfied with a good custom EQ on a headphone that is already close to something they enjoy (or again, just something clean, extended and fairly smooth, so EQ is physically possible without bad side effects).
-6/ And if none of that works for you, then and only then, would I go the traditional audiophile way of no EQ and swapping gears like crazy while praying to get a lucky combo that you like more. Give up on flat accurate sound reproduction and just go with what your guts tell you, as that's the only hope with that non fidelity method.


And I still ended up writing a book... I started telling myself I would only discuss FR to make it short. I did not do that.