What you said sounds a bit complicated, I wasn’t able to fully understand it.
It turns out “neutral” is a very complex term, not well defined term. Neutral should mean that it sounds like a pair of studio monitors - but this sounds impossible to achieve with headphones. How are you going to recreate the sound of something that is so bigger than the drivers from headphones? What speakers can be used as a standard for “reference”? Too many factors that are simply to complex to make a formula for “neutral”
Speakers that have a flat frequency response in an anechoic chamber have stopped being so useful, because the large majority of engineers or listeners will never be listening to those speakers in an anechoic chamber, so the “perfect” tuning would sound much different in your home - this is similar to HRTF with headphones. The headphone has this “great” frequency response and tuning that the manufacturer made with the help of professional measurement rigs, but then you remember that it will be a human ear that is listening to the headphones, and that frequency response get altered (HRTF)… that’s why I mention calibration/tuning that is done on a personal level. It’s not something that I will or want to do, I am creating an ideal scenario to better understand this topic.
Ignore the rant, focus on what I am about to say. Uncolored is a term that should be dependent on the frequency response, therefore it should be achievable to make an uncolored headphone. Ollo achieved this on the measurement rig, however, it will get altered because of the HRTF, so you won’t be hearing the same flat frequency response that the measurement rig made. So, the ideal situation:
You want the headphones to actually sound and have a flat frequency response to our ears, this is what you want the user to hear. The only way to do that is to make molds of your ears, several data and measurements of and from your ears, measure your HRTF, and then tune the headphones according to these molds (you would have mics in the mold ear). This would let you to tune any headphones exactly to your ear, so the frequency response that is being made with the help of the mics and software will be the closest to what you will be hearing. This would allow you to aim for a flat frequency response.
My question is whether this personally calibrated flat frequency response would be the definition of “uncolored”? It would be the most balanced, right? Since “no” frequencies stand out?
Would it be neutral? Probably not, at least from my understanding of “neutral”. I am not sure about this, so if you have an opinion about this, please share it:
Would a headphone with a flat frequency response (according to your HRTF and other information about how your ears change the sound) be able to be considered “neutral”?
I am specifically and only talking about headphones. So you don’t have sound bouncing off of your body, face, etc. The headphones are around your ears, so it would be the pinnae and other parts inside your ear that you would take into consideration - it’s different than with loudspeakers, where you have several different factors that play a major role.
To continue my rent from above. Essentially any frequency response that has been made with the help of a measurement rig (like G.R.A.S.) will be inaccurate as to the frequency response that the listener will actually be hearing (HRTF). This means that you cannot create a universal frequency response that will match the HRTF of the masses.
HRTF is the change of the sound source’s frequency profile by the unique characteristics of the listener
A head-related transfer function (HRTF), also sometimes known as the anatomical transfer function (ATF), is a response that characterizes how an ear receives a sound from a point in space. As sound strikes the listener, the size and shape of the head, ears, ear canal, density of the head, size and shape of nasal and oral cavities, all transform the sound and affect how it is perceived, boosting some frequencies and attenuating others. Generally speaking, the HRTF boosts frequencies from 2–5 kHz with a primary resonance of +17 dB at 2,700 Hz. But the response curve is more complex than a single bump, affects a broad frequency spectrum, and varies significantly from person to person
That is my point. The ideal scenario to be able to achieve a certain quality or frequency response would be to make the headphones and tune them to the individual’s ears. This avoids the argument of “this will not sound like that to everybody, thus the point is incorrect”. This lets the argument be focused on the quality itself, like:
Is a flat frequency response the definition of “uncolored”, “reference”, or “neutral”?
This question would refer to the custom tuned headphone that will let the listener actually hear the tuned frequency response, unlike with the case of mass production headphones (every headphone on the market) that have been tuned on a measurement rig (aritificial/perfect), and not on human ears. Since the HRTF significantly varies from person to person, the manufacturer’s frequency response that wasn’t done on human ears will significantly vary from person to person.
So, I decided to look at it from another perspective, and that perspective is that the headphone is being tuned for you specifically, and that the headphones will only be listened by you. This isn’t being done, nor do you have a company that is doing this (to my knowledge) - from a business point of view, this would be a very expensive business and product. I do not want to do this, I just want to look at things from this perspective theoretically - just so we are free of the vastly different HRTF bias that makes things impossible to be universal (impossible that the quality or frequency response sounds the same to everybody).
I don’t know if I made it more clear. Let me know if you want me to elaborate, or ask a specific question so you understand what I am trying to say.
EDIT: Look: https://youtu.be/a-JGAobDwGs?t=630
This is the idea. With this setup you know various things, can measure your own HRTF, and by knowing that you can tune the headphones specifically to your ears and data from your ears, this lets you achieve the desired frequency response.
This video is another way of discussing live recording methods, and the 1960s Bose 901 direct-reflecting technology. Top-tier recording engineers have done it for decades at the master recording level – the get the sound close to ‘intended’ or ‘good’ for everyone who listens to the recording. It’s shifting the process from the microphone to the headphone/playback stage.
I’m not sure what that gains on practical basis (recorded sources with personal equipment) over simply listening and adjusting EQ knobs or sliders as desired. Or, using DSP in a preamp to add reverb or hall effects at playback. One can either tweak the source or focus on the individual, but the perceived outcome may be ‘similar’ or ‘identical’ pending testing.
Each body is very different, but a large part of childhood involves learning and relearning how one’s own sensory organs function as they develop and grow. The brain will create one configuration then dispose of it and create another similar configuration, and so forth. As an adult one has a cohesive organs+brain functional configuration that continues to change and drift at a slower speed. Some functions decline with age, and the fall-off rates vary by body subsystem and by individual.
With the video method, measurement ambiguity follows from test error for human anatomy (i.e., limitations to measurement without destruction), rig material differences from human tissues, training/learning to unconsciously compensate for the environment which shifts perception over time – aka habituation (i.e., whatever you are using starts to sound more normal), fatigue (i.e., neural receptors function differently with extended use), bone conduction (i.e., different bone density), pads and seal (i.e., impact of head shape, density, reflection), and probably stuff I can’t think of right now too.
Bottom line for me: there are many ways to skin this cat.
In my opinion yes. Flat frequency response of sound exciting my entire sound perception anatomy (ear drum, inner ear, brain, etc) and perceived as neutral are the same. If you disagree we are using different definitions of “neutral” I think.
Would I personally prefer a neutral tuning to any other tuning? Not necessarily. I would probably add a bit of bass boost, especially if it is a studio recording as I wouldn’t be as concerned about being faithful to the live experience. But any changes I would make to color the sound would be applied to the neutral tuning as a reference since that tuning eliminates any biases created by the headphones. The perfectly neutral tuning gives me a baseline to work from.
Yes I’ve watched that entire video months ago. Here is how he puts that theory to practice:
I should probably note that I have not actually tried this - yet. But I intend to. I suspect I will get similar results to using the Harman target since they are similar ideas except the Harman is not personalized to me. It will be interesting to see if I can hear a distinct difference and if so, decide which one I prefer.
Another method I’m looking into is using this audio plugin:
I personally thought that this was correct, that this was the way to achieve a “neutral” and “reference” headphone, but the folks over at Hifiguides think otherwise - and they stated their strong points. They have a very valid point as to why this isn’t neutral.
I am currently in the place where I am trying to take several opinions and come to a single sound conclusion. To me it made sense that a flat frequency that a person would actually hear is what “neutral” is. If you were to produce a pure tone, it should sound like that to the listener. Isn’t this neutral?
Some say that “neutral” should try to replicate a speaker in a well treated room (something along this lines), but then… headphones will never achieve the speaker quality, so why try to replicate something that is impossible to replicate? This is my argument against that. Not only is the speaker driver several times larger than the headphone driver, but also - headphones are on your head, you just hear them through your ears. Speakers are in front of you (usually), you feel them all over your body, you feel them in your chest- you cannot replicate this with a headphone. When I said “flat on its own”, I was referring to the concept of having a headphone that actually had a flat frequency response to your ears, a concept Nick disagreed with. I think some may doubt this concept of individually tuning headphones to a persons ear and achieving a “neutral” headphone because it hasn’t been done by companies. Somebody can say “Oh, Voja, why do you think you are so smart trying to make a research breakthrough? If this was true, we would’ve already seen companies do this”. I look at it from a different perspective:
1˚ This method of making a headphone product for individuals would be extremely expensive, and not profitable enough to be made into a large scale business, this is why no company is doing it. You need to make molds for each individual, then throw these molds away. Expensive. Takes a lot of time and resources. But it is not unachievable.
So shouldn’t a neutral headphone be uncolored and produce tones as they really and truly are (notice I said “are” instead of “sound”, this is because “sound” is subjective to our ears)?
I am not the only one who has defined the term “neutral” as a flat perceived frequency response (no stand out frequencies), meaning it requires individual calibration to be achieved. But there are valid arguments that are definitely worth taking into consideration. I will quote some replies that I think are worth discussing, they stated some good points as to why “neutral” is not a perceived flat frequency response:
I think you’re simply refusing the notion that “flat” for headphones is not objective or universal and it’s not a graph that looks anything like a straight line
The point of “neutral” is to not sound different from reality. To achieve that, headphones have to have a response with hills and valleys, not a straight line. The hills and valleys will be different for each headphone because their drivers and cups are also doing different things to the sound, and also different for each user because their ears will be doing different things to the sound until it reaches the eardrum. What will natural FR look like on a headphone? A curve with hills and valleys, not a straight line. I don’t know how else to answer this.
So, those are the two points.
But “not sound different from reality” is natural (not neutral), isn’t it ?
If the second argument is true, than it would be impossible to define “neutral”. Because you would need to define these hills and valleys, you would need to define “reality”.
How can a human ear determine what neutral and flat truly is?
To which I got an answer:
The listener compares the sound of the headphones to the sound of anechoically-flat-responding speakers when used in a good listening room, which is how Olive & Welti determined the Harman target and also how dr. Griesinger’s equalization method works. “Flat” means it should sound like listening to natural sources that are away from your body and not attached to your head, so that’s what you compare to. This is not an abstract mathematical question, it’s a question of two types of human experiences and how you make one as similar as possible to the other.
Which would mean that there are many subjective factors. You cannot pin point what speaker you are trying to replicate, at what position, what room, and many other factors. This definition of “neutral” would be impossible to achieve or define.
I have concluded that “universal neutral” is an impossible concept. This is simply due to the fact that hearing and perception of sound is subjective (HRTF ), but you can make this concept possible if your aim is to make a “subjective objective” that would be “true” only to the listener’s ears. This is what I mean by personal calibration/tuning. You can achieve a flat frequency with this personal accurate tuning of the headphones, but would this flat frequency be considered “neutral” (not natural!!!)?
I think most people gave up on the concept of “neutral” because of HRTF, because of the fact that our ears change the sound, and that we have our own “personal frequency response” (HRTF). But if you make it possible by personally tuning the headphones for the user, you are not trying to say that these personalized tuned headphones are the definition of “neutral” for everybody, but to your ears specifically. Since our hearing of sound is subjective, there is no point in trying to achieve a “universal neutral/flat” because every person will hear it vastly differently - meaning it is not objective. Which brings me back to my point for frequency responses for audiophile headphones, they don’t tell you how this headphone will sound for you - everyone’s HRTF is different, meaning that the frequency response that the microhpones “heard” from the measurement rig is not what your ears will hear. BUT, if you make headphones and take several factors from our ears into consideration, then you “know” how this headphone will sound to your ears, meaning that you can achieve things such as a flat frequency response.
It is a very complex subject, hence why I write my replies so wrong. People don’t seem to be accepting the idea of being able to manipulate how the headphone will actually sound to somebody… it is a possible concept, and it is a concept that David Griesinger talked about in the video I attached. A lot of work, time, and money would be needed, but we are trying to form a conclusion here.
It’s not as though we are trying to agree that “X” headphone sounds like “x”. We all hear differently so there will always be an argument. But if you can calibrate a headphone to your own ears (not using EQ, since if EQ was that efficient, you would technically be able to calibrate the headphones to any desired frequency response), then you could achieve specific and accurate frequency responses that are subjective to your ears and your ears only. Objective sound in headphones is an impossible concept.
I am persistent about making a conclusion. I really think I am getting closer, but the debate of “neutral” and what is “neutral” is something that I look forward to. I want to know what can be considered a “neutral” headphone on a personal level (meaning to someobody’s ears). I already confirmed that an “uncolored” headphone would be the one that sounds flat to the listeners ears (requiring the personal calibration method I talked about in several of my replies above)
A few points to help you come to your own conclusions:
Recreating the tonality of speakers using headphones is not the same as attempting to recreate the entire experience of listening to music through speakers with headphones. The former is hard to do perfectly but easy to do well using EQ. The latter is hard to do period and requires EQ combined with sophisticated DSP, head tracking, etc.
Earlier in this topic you can see Resolve theorize that individual differences in HRTF may not make as big of a difference in perceived headphone tonality as many think. Things as simple as how the headphones are positioned make similar levels of change to the FR yet are barely noticed. So an average across many listeners may actually be close enough to how you hear tonality.
Two headphones with identical FR will have the same tonality. That is not to say that two headphones with identical measured FR will have the same tonality.
Using the term flat frequency response implies you are looking at a FR graph that is a straight horizontal line. However such graphs are a measurement and are often compensated to a particular target curve anyway. This may be where some of the confusion between flat and neutral comes from.
There are very good reasons why speakers in a room is a valid reference to use for determining neutral sounding headphones. For one thing we know how to measure if a speaker is neutral. For another the recording engineer is using speakers in a room and we are trying to replicate what they heard.
At the end of the day I don’t think anyone would argue that it would not improve the sound of headphones if they were calibrated to the individual listener. The argument is how to do so.
And BTW it’s not like there isn’t any interest in this from large manufacturers.
How would you define tonality? Are you referring to the frequency response?
If that is the case, then it should be incorrect. If HRTF means the altered frequency response (frequency response at the ear-drum), then it should be firmly incorrect. David Griesinger did several studies on this matter, and it can also be seen in the video I attached. While positioning of the headphones certainly matters, I think it would be impossible to position the headphones the same on every subject. Then you not only have the differences in our ears (which there are many) but also the head shape and size. I personally believe that the ear structure significantly affects the original sound. Or perhaps I misunderstood Resolve
I found this to be the point that is sticking out the most. Can you elaborate? I like how this sounds, but cannot really put it in perspective.
Yes. You mean to say that neutrality has much more to do than just a FR graph, correct? Then I could put the argument that even if a headphone was individually tuned to your ear to have a flat frequency response, it wouldn’t be eligible to be called “neutral”, correct?
If this is by the anechoic chamber method, then there is a strong argument. If you are referring to something else, please specify what method of measuring speaker’s neutrality you are referring to.
I believe David Griesinger did that, I attached it not too long ago and time stamped it. Also here is a quote from one of his studies:
Dummy head recordings from heads with anthropromorphic pinna can give good results if the head is properly equalized
and headphones can be matched to an individual listener
Finding the correct equalization for the dummy can be difficult – but can sometimes be done by spectral analysis post-recording.
All available dummy head models will give inaccurate results when used to equalize headphones.
Headphones can be accurately equalized for a particular listener using eardrum pressure measurements with probe microphones.
The first two are a must read. But David’s craze was about binaural recordings, while mine is:
1.The definition of “neural”, i.e. how would a frequency response (at the eardrum) look like for a “neutral” headphone?
2.What does a flat frequency response (at the eardrum) mean? Is it neutral?
What I was able to confirm (thanks to your attribution) is that a flat frequency response (at the ear drum) would be the definition of an uncolored headphone. This headphone would have to be calibrated to your ear with the help of a cast of your ear (this would basically replicate your own whole ear structure), your HRTF, your eardrum pressure, and possibly other factors that I am not aware of (I am no scientist, just a guy using the internet to obtain this knowledge).
What I personally believe, but haven’t confirmed:
Since Ollo Audio S4X did measure almost perfectly flat in the frequency response (on my friends earless measurement system), this would probably make it the most uncolored headphone on the market.
This means that the individual frequency response (at the eardrum) would be the closest to this “ideal idea” of a flat frequency (at the eardrum) - it would be the least distorted and changed version. Since we don’t have manufacturers individually calibrating headphone to the listener’s ears, this would be (what I consider) the closest we will get to this ideal situation (without individual calibration to the listener’s ears; something that David stated in his studies is crucial in order for headphones to be accurate to our own ears). We already know that the frequency response (measured at the measurement system - microphone setup and software) for all of the headphones on the market currently, will get changed by the time it reaches our eardrum, thus any frequency response from a measurement system that steps away from being flat will have more coloration/distortion, this is the reason why I would consider calling the S4X the most uncolored headphone on the market - I don’t think it currently gets any closer than this.
Would this make it the most neutral headphone on the market? I don’t know because I have yet to confirm the definition of neutrality - but I strongly believe it is much more than just FR… which would mean that I will avoid this statement since I cannot measure things such as soundstage (etc.) so there is no point, I cannot make a sound hypothesis - which is the only thing that I am trying to do.
For the sake of this thread and people who don’t want to bother and take the time to read my long answers (completely understandable), I want to share the following links and studies from David Griesinger (his website isn’t online, so I am sharing the cached versions of his studies, most of them were presentations so pictures are missing):
I strongly recommend reading the first two links.
1.Frequency response adaptation in binaural hearing
2.Binaural Hearing, Ear Canals, and Headphone Equalization David Griesinger
3.The necessity of headphone equalization
4.Recent concert hall research findings and a method to equalize headphones to an individual at the eardrum
You will notice that the Harman team was involved in some of these studies as well
Some may argue that the FR of headphones does not describe such things as detail, clarity, punch, soundstage, imaging, locality, etc. But we all agree that it does describe tone. So I am using the word tonality to describe this aspect of FR.
HRTF most definitely affects tonality. What I am saying is that listener’s individual HRTF may not vary enough to affect perceived tonality. So an average HRTF compensation is good enough to make you perceive a headphone as neutral without compensating for your individual HRTF.
If two headphones are vibrating your ear drum in the same way, they will have the same FR and sound exactly the same (including tonality). But current headphone FR measurement techniques are not accurate enough to perfectly capture this.
It may be possible to measure FR at the ear drum accurately enough (levels) and with enough resolution (number of frequency bands measured) to perfectly capture tonality though - remember I’m using tonality as just part of what FR contains.
I was trying to make clarifications without stating opinion. So all I meant was to keep in mind what a reader might be thinking when you say flat.
A flat graph can show neutrality for speakers, but the limitations of measuring headphones means that merely saying a headphone measures flat does not mean much for neutrality. For instance you have to how the measurement was taken and whether any target curve compensation has been added.
Yes, exactly. It is pretty well accepted that if a speaker measures flat from 1m on axis in an anechoic chamber and smoothy transitions off axis then it will sound neutral in a good room.
My bad. Griesinger did do that. I should have said the argument is how BEST to do so. Griesinger’s method has two big issues imho. One is that it relies on the listener being able to accurately distinguish tone levels. But a bigger issue for me is that it isn’t finely enough measured. The difference in frequency between the tones is too wide to create a high resolution compensation EQ. At least higher resolution than you’d get by using non-personalized version like Harman.
This is very interesting because I was not aware of anyone using probe microphones for this. I assumed that was because this technique was not accurate enough or didn’t have high enough resolution. @OldAudioGuy should take a look at this as he mentioned using it himself in another topic.
I’ll take a look, thanks. His research around binaural recording is one reason why I said
Griesinger is attempting to not only correct the tonality of headphones, but also the soundstage whereas Olive was only concerned about tonality. Too often I’ve seen this concept confused in discussions.
No. An earless rig cannot accurately simulate the FR that would occur at the ear drum in a human, even with compensation. I’m guessing you are thinking that since speakers play the room and headphones play your ears, that an anechoic room is to speakers as an earless measurement rig is to headphones, but it doesn’t work that way.
Yes. This is why I am avoiding the term “neutral”. I cannot form any type of hypothesis or theory based on something that simply cannot be measured. This is the reason why I am at least trying to prove something with the FR.
Hm. You are basically saying that if a headphone was tuned on a dummy head setup with an artificial pinna (like G.R.A.S’s KEMAR measurement system), that it will be accurate enough to ensure that it sounds close to the tuned frequency response on this measurement setup? In other words the tuned frequency response from this setup will sound mostly accurate at the eardrum of most people?
This is something that I question, since people perceived the pink noise quite differently, at least differently enough that it is audible that it does not sound like the original sound (David’s video).
The argument is that many people started disapproving this. If the speaker measures flat in an anechoic chamber, it will not measure flat at your home (several reflections, etc. - you would have to treat it acoustically pretty well). Some people’s argument is “Why do I care if it measures flat in an anechoic chamber, I will not be listening to the speakers in the anechoic chamber”. But in the similar way, you cannot create a product for the masses and also create it exactly for their setup and home - this require an in-person service that would calibrate the product exactly to your environment (speaking about speakers).
I think this is something that was done later on in the video. Here is the time stamp that I am talking about: https://youtu.be/a-JGAobDwGs?t=628
Watch it right about the 10:48 minute mark.
Hear my hypothesis out - if you could use your own dummy head that has your ears (accurate casting of your ear structure), then you will be tuning the headphones to your ears instead of KEMAR’s (because you are not KEMAR). If you have the microphones placed inside your very own dummy head, then you are basically getting as close as I think you can get to individually tuning/calibrating headphones to an individual.
I noticed this. I believe he did back up his methods with the fact that the subjects could repeat the result consistently, which would “prove” that his method is effective. However, I am only concerned about the FR/tonality, since that is just about as much as we can “accurately” measure.
Yes, an earless rig is not even trying to simulate what happens at the eardrum. This is why I think it is the most “objective” method. Because you are tuning the headphones without the human factor in it, which would leave the alteration of the “true” frequency response (human factor) to vary from person to person, but it should never stray too far away from the frequency response on the earless rig - the way I am thinking is that it will never become a V-shape or something of that sort. It should be the least distorted way to achieve an objective FR (to be objective you usually leave the subjective factors out).
However, perhaps I am wrong and I do not understand the concept well enough, in which case I would greatly appreciate if you can focus your answer on explaining this.
I am not sure if I thought like that. The way I am thinking is that the earless rig leaves the biggest obstacle and subjective factor out - our ear. Since the majority of the listeners will not be KEMAR, they will hear the headphones differently. I think that the frequency at the eardrum from a pair of KEMAR tuned headphones would vary more than it would with an earless rig - in other words, I think that the frequency at the eardrum from KEMAR tuned headphones will not be as close to the OG FR as it would on the earless rig. I am thinking that it is more objective with an earless rig, but it may be the completely opposite, and that actually for the very reason that the subjective factor is present on KEMAR setup, that it is more accurate and objective.
Hi thanks for the mention, sorry it’s a long thread so haven’t read it all so apologies in advance. Probe microphones are in common use for hearing aid target matching but not for headphones at the moment. They are fiddly to set up for accurate high frequency measurements (10k) they need to be right on the tympanic membrane (1mm). I think this is more for research at the moment than a practical solution for equalisation or HRTF measurement though that might change. Suprathreshold loudness balancing as used by Griesinger will give your amplitude response though not the group or phase delay for which microphones are required though you could use existing normative data for this. Other suprathreshold and threshold techniques are available, high resolution can be achieved but takes longer though some shortcuts exist.
I think headphone manufacturers have looked at all this but decided it mostly wasn’t worth it at the moment, maybe it would be a bit disruptive for the high end market as any cheap but reasonably linear low distortion headphones could be accurately corrected to an optimal response with suitable filters - hybrid head tracked linear phase filters for full HRTF 3D sound field simulation or just a minimum phase filter for straightforward EQ. This has happened to some extent for enthusiasts with Autoeq etc. Some manufacturers have started to rely on EQ to make the products work now I think Audeze does for some models. The EQ is just from population data though so individual listeners will have different percepts.
I am happy just using Equaliser APO for minimum phase IIR EQ and Waves NX for the head tracked linear phase HRTF simulation. I seem to be almost alone in liking this though.
Agreed as long as the artificial ear you are using is also made out of something that mimics the acoustic impedance of a real one. IE - Don’t use plaster for the pinnae. Use something flexible like a real ear.
Another option that would be accurate enough to use for tonality correction would be a Gras unit with your own ear shape instead of the stock one.
Objective is not the same as accurate. I think if you used such a rig and applied a proper compensation to the raw measurement and the FR came out flat overall, you could probably say the headphone is neutral. BUT the frequency peaks and valleys measured would be less and less accurate as you go higher in frequency. So you could describe the overall tonality as neutral but you’d probably have an awful sounding headphone.
In the end I think you feel that the individual variation in HRTF has more of an influence on perceived tonality than it actually does. Griesinger was trying to turn a stereo recording meant to be played over speakers into an accurate simulation of actually being there, or at the very least a simulation of a binaural recording. This requires much more accuracy and personalization to pull off than just making a headphone sound tonally neutral.
If you can accept that individual ear variation is not nearly as important in how we perceive tone, then I think you’ll agree it is far better to use a full simulation of an average ear for measurements than any earless rig. And indeed that is what anyone who has the means to purchase something like the Gras is doing.
Here is what Griesinger did (quoting from one of his studies):
Pinna and ear canal casting:
Pinna and ear canal are filled with a water-based alginate gel. The resulting mold is immediately covered with vacuum degassed silicone to produce a positive cast.
More on casting:
The silicon material was “Dragon-Skin” from Smooth-On with hardness of Shore 10.
The cured silicon positives are covered with more silicon to produce a durable negative for further reproduction.
The outside surface of the silicon pinna are cut away with a small scissors to reproduce the compliance of a real pinna, which varies from shore 3-10.
Tiny probe microphones are attached to the apex of the eardrum cavity, and a resistance tube of about 3m in length is attached to the center of the eardrum to simulate the eardrum resistance. 18 gage PVC was used.
The probe microphones were calibrated to be flat to about 14kHz as referenced to a B&K 4133.
DSP is used on the microphone outputs to apply the resulting equalization.
The result matches probe measurements of my own ears within about 2dB.
Paraffin wax is used to fill the space inside the head around the ear canal and resistance tube to eliminate microphonics.
The outer head was cast with a high-density artist’s foam material from Smooth-On. This material is easily formed and cut.
I do think this is the case. But I cannot confirm nor deny to what level our HRTF differs. I based my strong opinion on the pink noise test in David’s video. It just sounded very different, and if I was to put that on the scale of music listening, I just cannot imagine how much differently something would sound.
I have the Ollo Audio S4X, and I didn’t find it to sound awful. But very noticeable much different than the other headphones I own. Interestingly enough, it first hit me how different and “raw” it sound when I hopped in a game (CS:GO), and shot a sniper rifle. My ears immediately noticed that it is much different than usual sound they are used to. I personally am not a sound engineer, so I cannot tell if a headphone is neutral or not, I just noticed that the treble seemed more revealing, more “raw”, perhaps easier for the ear to notice imperfections.
Ollo Audio used the “G.R.A.S 45CC using IEC 60318-1 ear simulator and Dewesoft AD conversion Sirius system”. It is an earless measurement rig, however, it is simulating the eardrum (from what I understand). According to Ollo, G.R.A.S. considers this the best setup for headphone testing, but I do know some manufacturers use the KEMAR setup from G.R.A.S. Now, I am not sure which one is more accurate and useful - G.R.A.S. has their team of engineers and researchers behind their products, I’m pretty certain they know what they are doing, and am pretty sure each rig has their own use.
I still have not settled on a single statement regarding Ollo Audio’s S4X headphone, since you did bring up more valid points in the reply from before, more specifically:
Maybe you would change the wording when we take into consideration a measurement system from G.R.A.S. which actually does try to simulate the humans eardrum? Or do you still believe that a setup like the KEMAR would be more accurate and more suitable to achieve this type of result?
I think Ollo Audio’s team did a great job at explaining their process of tuning the S4X, I would pay closer attention to their article: https://olloaudio.com/pages/measurements
(please let me know if the page is in English! - if it is not, change the language to english and click my link again)
They also made two videos… but they feel more like a reality show, either way, if somebody wants to help and contribute, give them a watch:
Mr. Rok brings up several good points, and he is also aware of the challenges (at one point he brings up two very good points: that there is no standard to mimic your ears and your taste). In the article he basically goes over everything that his research team put their work into. To me it seems like they put in a lot of thought into their product, so I am just trying to understand how this would translate to the listener and what it all means to the listener.
I may be thinking old information, but as I recall anecholic chambers typically only damp frequencies above 200hz. Back in the 1972 when my Rectilinear III speakers were being modded by the then owner, he took them to the anecholic chamber to test. He was a grad student at Penn State in engineering acoustics, who returned to college after some work in the audio industry.