• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Master Complaint Thread About Headphone Measurements

I think this long write-up is beside the point. I understand Harman is a well-researched preference, I have no qualms about that. The point I was responding to initially was this statement by Amir:



The first bolded part is about that if the producing side of music has disagreeing opinions, then Amir states that they should do their own research. I fully agree with that.

The second bolded part, is not right. It doesn't make sense.

1) According to this statement, if the engineer has a very different preference than Harman, then apparently he should just suck it up and listen in a way he doesn't find enjoyable. I wonder how great recordings will turn out then.
2) It builds on a misunderstanding about the creation side of music:

All music is produced, mixed and mastered in comparison to its genre. The only time when you hear engineers say they use no references anymore, is when mastering engineers specifically say they've been using the same system exclusively for so long that they don't need it anymore. All engineers will always start with references, and basically only a very small subset grows skilled enough to not need it, and having been attuned to their system is mandatory for that.

In result: only if audio professionals tune their ears to a different system than the one they work on, will they create some sort of circle of confusion in their brain.
But to make an analogy: that would be the same as if an F1 driver is practicing for his race in a speedboat instead of their race car. If he botches his race because the car behaves differently, then that is fully a failure on the drivers part. Not the fault of the speedboat and racecar industry not agreeing on a single standard.
To simplify, instead of using the term "Harman" as for headphones, let us focus on "neutral speakers in a well-treated room". (Based on my understanding; I don't work in the industry.) If the consumer has such a neutral playback system, what might happen if the music producer recorded the sonic event with neutral microphones, but mixed and whatnot with speakers or a room that incurred a large 1 kHz to 5 kHz dip? Maybe those speakers may already be to their preference such that they didn't make any EQ changes to that region. That consumer with a neutral playback system then would certainly not be hearing what the music producer heard and preferred. Alternatively, maybe the original neutral recording did have too much ear gain which due to the choice of monitoring system that already attenuates that region led to such not be corrected for with EQ, causing that consumer to hear a bright sound. Alternatively, the music producer finds that they do need to EQ the 1 kHz to 5 kHz region up in order for the neutral recording to sound right to them, but instead of treating that as a compensation for their playback system, they ship that EQed result with the final product, causing the consumer to hear that EQ boost.

Matching recording, monitoring, and playback system frequency responses would theoretically avoid alterations to preference-based changes made by the music producer. At least for speakers, flat recording, monitoring, and playback should be the way to ensure transparency of spectral content through the music production and distribution chain and to best separate EQ alterations from the original sonic event. If the music producer has a preference deviating from neutral, then they will EQ the neutral playback to said sound, and if they intend for the consumer to hear that same tonality and the consumer also has a neutral playback system, then they will indeed also hear that EQed preference. What's the point of letting the music producer monitor or EQ such that the music sounds the absolute best to them on their system of choice if we in perpetuating the circle of confusion let consumers use playback systems that may present that music with a drastically different tonality from what the music producer found best?

But I again suppose it is different if the music producer were somehow one to be able to monitor their music with their "preferred" playback tonality but already "know their speakers'/rooms' sound" enough to only make the kinds of EQ adjustments that would be revealed as necessary by a neutral monitoring system.

Getting headphones to sound as close as possible to said standardized neutral playback system, and for as many people as possible, is a different story for which I advocate for personalized HRTF measurements.

Anyways, in an ideal world, I wouldn't have to listen through upwards of 100 different classical recordings of the same piece in search of one that sounds tonally correct if not real. Maybe standardizing recording, monitoring, and playback neutrality could greatly reduce that substantial variability of recording quality and allow me to focus on choosing my favourite interpretation. Then one would have the variable of micing and mixing techniques insofar as I know many recordings that could sound tonally excellent but may still suffer from having entire string sections mixed to or imaged from singular points instead of properly spread across a line of musicians as heard in the concert hall.
 
Last edited:
Your post is the textbook definition of the confusion that the creation of the term "circle of confusion" caused.

Let's start with a hypothetical situation where I am the artist and producer, and you are the listener. I am visiting your house. We both love rock music, and we happen to both love Iron Maiden. (Insert your genre and artist of choice). You let me listen to your headphones, that you are extatic about. Yet as it turns out, while we both love to blast 2 Minutes to Midnight, you love the bass a lot more than I do. The pressure of the bass is literally a bit uncomfortable to me, so I turn it down a little.

Harman headphones can be replaced with the Toole speakers in this example like you suggested ('straight in an anechoic room, placed in a semi-reflective room' I believe).

Now I am back at work, and today I am recording a Iron Maiden cover band. They want to record a cover of 2 Minutes to Midnight, true to the original.

My headphones don't have so much bass, as I don't prefer that, and I use them for pleasure listening too. I can record and mix the song in a way that sounds great to me, and tonally equal to the original. When I send it to you to listen, you listen on your headphones, and you hear a great recording, similar to the original you heard yesterday. You are happy. The artist is also happy because youre happy, and that is their true intention. I am also happy because I was able to translate.

However, if I had a standardised 'Harman' headphones (or Toole speaker+room), then I'd have to two choices:
1. Turn the bass down in the record for it to sound good to me. Negative effects: I lose the relation to the original Maiden song. And it will sound weak to you.
2. Keep the bass at the same level as in the reference original. Negative effect: I personally hate the sound in the mixing proces. I make bad decisions because continuously hate what I'm hearing, and I can't put my heart in it. Perhaps I can grit my teeth, power through and make something that is good enough.

Both choices will lead to nobody being happy.

The conclusion is that because preferences exist, it is detrimental to force everyone to the same standard. You will not get "what the artist intended", instead you'll get further away from it.

Even if the Harman research is completely and irrevokably correct* for everyone in the world, it doesn't disprove this. If 64% of people like a certain curve over the existing songs, then 36% prefer something else. Don't punish the 36%.

*I doubt that it is, because there are many competing curves that people are a fan of, especially for changes in the 1-5kHz area. There's also research done on country preferences and the results are not at all that everyone likes the same flat line in the room. But let's assume it's fully the truth.

To really drive the point home: as is indicated by the poll on this website, and my unofficial observations, producers fall more into the lower bass category (24%). If you, the consumer, really want to hear what the artists hear, then you'll have to live with their lower bass level preference. So in fact, this punishes 76% of the people.

Only if 100% of the people liked the same curve, based on the body of recorded music that we have now, would standardisation make sense.

Then, the circle of confusions 'solution', is that every studio, microphone, monitorsystem in the whole world must be standardised, as well as every living room (size and reflectivity), speakers and headphone. As well as every concert hall. Achieving world peace is probably easier.

To get exactly what the artist heard, you should even listen at the exact loudness that the artist intended, to avoid a difference in tonality due to the equal loudness contours (FletcherMunson/Phon).

The alternative is to keep doing what we do now: artists and consumers/listeners alike buy the systems that make the majority of their music sound great to them, and the translation happens automatically. It costs nothing.

The circle of confusion is something that sounds deeply profound, but doesn't hold a bit of truth in it.
 
Last edited:
To simplify, instead of using the term "Harman" as for headphones, let us focus on "neutral speakers in a well-treated room". (Based on my understanding; I don't work in the industry.) If the consumer has such a neutral playback system, what might happen if the music producer recorded the sonic event with neutral microphones, but mixed and whatnot with speakers or a room that incurred a large 1 kHz to 5 kHz dip? Maybe those speakers may already be to their preference such that they didn't make any EQ changes to that region. That consumer with a neutral playback system then would certainly not be hearing what the music producer heard and preferred. Alternatively, maybe the original neutral recording did have too much ear gain which due to the choice of monitoring system that already attenuates that region led to such not be corrected for with EQ, causing that consumer to hear a bright sound. Alternatively, the music producer finds that they do need to EQ the 1 kHz to 5 kHz region up in order for the neutral recording to sound right to them, but instead of treating that as a compensation for their playback system, they ship that EQed result with the final product, causing the consumer to hear that EQ boost.

Matching recording, monitoring, and playback system frequency responses would theoretically avoid alterations to preference-based changes made by the music producer. At least for speakers, flat recording, monitoring, and playback should be the way to ensure transparency of spectral content through the music production and distribution chain and to best separate EQ alterations from the original sonic event. If the music producer has a preference deviating from neutral, then they will EQ the neutral playback to said sound, and if they intend for the consumer to hear that same tonality and the consumer also has a neutral playback system, then they will indeed also hear that EQed preference. What's the point of letting the music producer monitor or EQ such that the music sounds the absolute best to them on their system of choice if we in perpetuating the circle of confusion let consumers use playback systems that may present that music with a drastically different tonality from what the music producer found best?

But I again suppose it is different if the music producer were somehow one to be able to monitor their music with their "preferred" playback tonality but already "know their speakers'/rooms' sound" enough to only make the kinds of EQ adjustments that would be revealed as necessary by a neutral monitoring system.

Getting headphones to sound as close as possible to said standardized neutral playback system, and for as many people as possible, is a different story for which I advocate for personalized HRTF measurements.

Anyways, in an ideal world, I wouldn't have to listen through upwards of 100 different classical recordings of the same piece in search of one that sounds tonally correct if not real. Maybe standardizing recording, monitoring, and playback neutrality could greatly reduce that substantial variability of recording quality and allow me to focus on choosing my favourite interpretation. Then one would have the variable of micing and mixing techniques insofar as I know many recordings that could sound tonally excellent but may still suffer from having entire string sections mixed to or imaged from singular points instead of properly spread across a line of musicians as heard in the concert hall.
I would bet there's loads of different ways classical music (orchestra) could be recorded and layed down on a track, I can quite imagine the degree of variability you could get from recording to recording. I would imagine it's a lot more tricky to get right than say more modern types of music that are designed from the ground up to be a 2 channel experience that creates the virtual sonic soundstage that the artist/engineer desires. I don't listen to a lot of classical, occasionally Mozart, and I just don't really find recording quality to be a big issue with the music I listen to - if you've got Anechoic Flat speakers or an optimised headphone EQ then everything sounds excellent to acceptable. (The Mozart I've got sounds fine too.) I can certainly imagine classical music to be more tricky.
 
I would bet there's loads of different ways classical music (orchestra) could be recorded and layed down on a track, I can quite imagine the degree of variability you could get from recording to recording. I would imagine it's a lot more tricky to get right than say more modern types of music that are designed from the ground up to be a 2 channel experience that creates the virtual sonic soundstage that the artist/engineer desires. I don't listen to a lot of classical, occasionally Mozart, and I just don't really find recording quality to be a big issue with the music I listen to - if you've got Anechoic Flat speakers or an optimised headphone EQ then everything sounds excellent to acceptable. (The Mozart I've got sounds fine too.) I can certainly imagine classical music to be more tricky.

Agreed. Even if you were to standardise the use of the most common Decca Tree method. Intersting read:

But even if you standardise the microphones used and the size of the Decca Tree, every orchestra has a different number of musicians, and a different distance between each, in different halls, so you'd never end up with a 'true' recoridng. Not to mention that most listeners don't float above the orchestra.

Then there also no consensus on what is best. There are numerous other techniques, including where a microphone is laid down flat on the floor, and it sounds surprisingly good.
 
Last edited:
Agreed. Even if you were to standardise the use of the most common Decca Tree method. Intersting read:

But even if you standardise the microphones used and the size of the Decca Tree, every orchestra has a different number of musicians, and a different distance between each, in different halls, so you'd never end up with a 'true' recoridng. Not to mention that most listeners don't float above the orchestra.

Then there also no consensus on what is best. There are numerous other techniques, including where a microphone is laid down flat on the floor, and it sounds surprisingly good.
Interesting read, thanks, didn't know about that. Looks like it can indeed be quite complicated to record an orchestra!
 
Your post is the textbook definition of the confusion that the creation of the term "circle of confusion" caused.

Let's start with a hypothetical situation where I am the artist and producer, and you are the listener. I am visiting your house. We both love rock music, and we happen to both love Iron Maiden. (Insert your genre and artist of choice). You let me listen to your headphones, that you are extatic about. Yet as it turns out, while we both love to blast 2 Minutes to Midnight, you love the bass a lot more than I do. The pressure of the bass is literally a bit uncomfortable to me, so I turn it down a little.

Harman headphones can be replaced with the Toole speakers in this example like you suggested ('straight in an anechoic room, placed in a semi-reflective room' I believe).

Now I am back at work, and today I am recording a Iron Maiden cover band. They want to record a cover of 2 Minutes to Midnight, true to the original.

My headphones don't have so much bass, as I don't prefer that, and I use them for pleasure listening too. I can record and mix the song in a way that sounds great to me, and tonally equal to the original. When I send it to you to listen, you listen on your headphones, and you hear a great recording, similar to the original you heard yesterday. You are happy. The artist is also happy because youre happy, and that is their true intention. I am also happy because I was able to translate.

However, if I had a standardised 'Harman' headphones (or Toole speaker+room), then I'd have to two choices:
1. Turn the bass down in the record for it to sound good to me. Negative effects: I lose the relation to the original Maiden song. And it will sound weak to you.
2. Keep the bass at the same level as in the reference original. Negative effect: I personally hate the sound in the mixing proces. I make bad decisions because continuously hate what I'm hearing, and I can't put my heart in it. Perhaps I can grit my teeth, power through and make something that is good enough.

Both choices will lead to nobody being happy.

The conclusion is that because preferences exist, it is detrimental to force everyone to the same standard. You will not get "what the artist intended", instead you'll get further away from it.

Even if the Harman research is completely and irrevokably correct* for everyone in the world, it doesn't disprove this. If 64% of people like a certain curve over the existing songs, then 36% prefer something else. Don't punish the 36%.

*I doubt that it is, because there are many competing curves that people are a fan of, especially for changes in the 1-5kHz area. There's also research done on country preferences and the results are not at all that everyone likes the same flat line in the room. But let's assume it's fully the truth.

To really drive the point home: as is indicated by the poll on this website, and my unofficial observations, producers fall more into the lower bass category (24%). If you, the consumer, really want to hear what the artists hear, then you'll have to live with their lower bass level preference. So in fact, this punishes 76% of the people.

Only if 100% of the people liked the same curve, based on the body of recorded music that we have now, would standardisation make sense.

Then, the circle of confusions 'solution', is that every studio, microphone, monitorsystem in the whole world must be standardised, as well as every living room (size and reflectivity), speakers and headphone. As well as every concert hall. Achieving world peace is probably easier.

To get exactly what the artist heard, you should even listen at the exact loudness that the artist intended, to avoid a difference in tonality due to the equal loudness contours (FletcherMunson/Phon).

The alternative is to keep doing what we do now: artists and consumers/listeners alike buy the systems that make the majority of their music sound great to them, and the translation happens automatically. It costs nothing.

The circle of confusion is something that sounds deeply profound, but doesn't hold a bit of truth in it.
I would vie that your Iron Maiden case would be one of trying to reproduce the tonality of a recording that had already been afflicted by the "circle of confusion" if it doesn't sound right to you on a flat system, and is likewise constrained to a case-by-case coupling between your preferred playback versus a specific client's as opposed to a general population. Now, again, due to limitations of Harman headphones sans speaker crossfeed where I may indeed find differences in bass level preference (e.g. I may be wrong, but I think for your example, the range or perception of headphone bass levels as compared to flat speaker levels might be too variable, or the range of speaker in-room FR variations is probably quite different from the range of headphone FR variations; maybe someone could share a speaker model or room that actually in itself has a Harman-like bass boost or negative-gain shelf incurring the opposite) and ear gain (though such preferences may be rather well shaped by said "circle of confusion's" effect on recordings), I want to focus on the "circle of confusion" in regard to neutral speaker playback, and if ever headphones were involved, I would want to do probe microphone HRTF measurements to match the responses. My focus is on the production of new music (particularly acoustic), though indeed inherently thinking upon a currently "unrealistic utopia" of flat recording and playback everywhere. Transparency of loudness I suppose can be a real concern especially when the original live even really was earsplitting (e.g. Mahler Symphony No. 5), but at least the consumer has an easy choice and means of adjustment.

As for the concert hall or any live venue and its respective musicians and their instruments' unique timbres, these sound variations are in themselves what I would consider to be the ground truth that we may wish to transparently convey to consumers in hope of producing a "real" experience, though yes, it is possible for one including myself to sit in a concert hall and whether due to my choice of seat or the priming of the recordings I had listened to and the imperfect tonality of my playback system among other psychological factors to just not like the sound of the strings in that particular performance. In that regard, perhaps I am indeed impractically hoping for a world, at least a subset of audio perhaps specifically for acoustic music, where reality is held in higher regard than "preference"; i.e. it can be debated whether the real, live sound is necessarily always the preferred one, and if it is, then the question is of how to reliably relay that live sound which in my opinion demands the ending of the circle of confusion through mass standardization. First transport the consumer into the original venue, then let them adjust as they wish. Now yes, amplified music is probably a case that can sound considerably worse live than in recordings, or perhaps the studio recording method does inherently require special processing to create the illusion of an idealized "live" performance; for electronic music, "anything goes" in regard to the music's tonality, or the ability to hear the producer's "intent" would be even more important. But if that studio were calibrated to a standardized flat recording and playback, at least consumers wouldn't have to guess what kind of system would be needed to hear that original intent should they want to. I suppose then that my wish is rather to standardize the neutrality of music production while still letting consumers run free in regard to their preferences. If one had a huge sample size of neutrally recorded music, then the demographic preference curves would change accordingly and playback products could be tailored for such. At least for all that new "neutrally produced" music, those who desire will have confidence in how to hear the "original sound".

There have been some tonal issues with strings that I couldn't figure out how to fix with EQ, whereby I suppose I simply didn't like the timbre of that orchestra or ensemble, or the recording chain really was picking up distortion, or the microphone placements really were per the instruments' directivity picking up a harmonic distribution different from what the audience hears, like in that case of how you aren't listening from right above the orchestra, or that some mics' position might rather cater to listeners in the mezzanine or upper balcony from which I find the orchestra to be rendered too quiet. I have yet to come across an ambisonic symphonic recording made from a centered orchestra level seat. As for said ambisonic reproduction, I would hope for it to capture cases of my primarily hearing the direct sound of the strings while only certain brass, reed, or percussion instruments can be heard imaging from ceiling or rear wall reflections.
 
Last edited:
That's a very thoughtful post and I think I agree on most if not all on it. So if I may address your main point, the last sentence of the 2nd paragraph:

I suppose then that my wish is rather to standardize the neutrality of music production while still letting consumers run free in regard to their preferences. If one had a huge sample size of neutrally recorded music, then the demographic preference curves would change accordingly and playback products could be tailored for such. At least for all that new "neutrally produced" music, those who desire will have confidence in how to hear the "original sound".

This makes sense. There's something nice about knowing you're listening to music 'the right way / neutral / as intended', or 'as real as possible'. However, my thoughts on this:

As you mentioned, the seating position in the concert hall will massively influence the sounds. Research shows that listener preferences aren't equal, 2/3's of people prefer the more clear sound in the frond of the hall, 1/3 prefers the more immersive sound (more reflections, more bass, less clarity) of sitting near the back wall (this effect is stronger in certain designs than others). But which one is the intended way, or the most neutral?

You could ask the artist. However, If we imagine an acoustic duet of a double bass player and a transverse flautist, who is the artist? The bass player, the flautist or the original composer? If we say one of the players: while playing, the bass player will hear the bass the most, and the flautist will hear the flute the most. Both have no real idea what they sound like in the hall, let alone which one of the seats corresponds best with what they intended. They can never know. Unless they give their instrument to someone else, but even then there will be differences in playing, especially if these people are the best in their field.

But okay, let's make it not too difficult, so let's say all seats are equally intended, as long as the recording is true to the experience to a particular seat.

On a positive note: most acoustic music is already neutral. Most microphones used are pretty flat, and a mixing engineer wouldn't start doing heavy EQ-ing, because it'll change the timbre of many instruments in unrealistic ways. If you have a flute and a violin on the same recording, adjusting the fundamental note of the flute will also adjust the higher harmonics of the strings. There's little freedom there. The difficulty of using 3 microphones to record a 60-people orchestra is that you have very little control after the recording. This type of music all depends on the recording engineer.

Thus, if you listen on Dr. Toole's speakers (flat and smooth anechoic with smooth directivity in a semi reflective room), you are already experiencing as much of reality as you can. However, your room is not 'standard'. There's no one standard for how downsloping the room tilt is ideal (afaik). Unless you have also a (utopian) certified standard room, true ultimate reality is not attainable, at the very least you must set your tilt yourself. And even then, that room wouldn't be "most real", it would just correspond with the preference of most people. Since the original recording had nothing to do with your room, anything your room does is always unnatural.

Furthermore, stereo speakers can't reproduce the surround experience of a concert hall.
I know you know (as someone who visits halls), but out of interest see for example how much is going on in a concert hall: https://www.pnas.org/doi/full/10.1073/pnas.1319976111

So if you want ultimate realism, stereo speaker will always fall short. Binaural n-ear is maybe the best bet, but I suspect you'll hate the massive amount of reverb/reflections. In the real world you move your head and your brain gets cues to understand and enjoy the sound in that room, but you won't with in-ears. So while I can see the point of more realism equals more enjoyment, even if the sound is not 'perfect' , for these shortcomings I don't think it's a worthwhile goal to pursue on stereo or regular headphones / inears. But perhaps a spatial audio recording of a orchestra in a concert hall seat, played on either Atmos type surround or spatial audio in-ears with headtracking can get very close.

You mentioned string timbre. Recorded strings with a 'neutral/flat small diaphragm condenser (SDC) mic sound horrible, not at all as in reality. That's why violins are typically recorded with ribbon mics or vintage tube mics, known for their flattering smooth transient response. Many violinists also own a ribbon mic to make sure they can sound good wherever they perform. I have the pleasure of closely knowing two professional violinists. Bat ears, truly, and they insist that only ribbon mics can capture 'their sound'. This mic is objectively less accurate but subjectively more accurate. Why? I don't know. It doesn't really make sense to me in a logical way. Perhaps SDC's are exaggeratingly fast, and the ribbon mic's slow inaccuracy is still closer to the truth.
 
There have been some tonal issues with strings that I couldn't figure out how to fix with EQ, whereby I suppose I simply didn't like the timbre of that orchestra or ensemble, or the recording chain really was picking up distortion, or the microphone placements really were per the instruments' directivity picking up a harmonic distribution different from what the audience hears, like in that case of how you aren't listening from right above the orchestra, or that some mics' position might rather cater to listeners in the mezzanine or upper balcony from which I find the orchestra to be rendered too quiet. I have yet to come across an ambisonic symphonic recording made from a centered orchestra level seat. As for said ambisonic reproduction, I would hope for it to capture cases of my primarily hearing the direct sound of the strings while only certain brass, reed, or percussion instruments can be heard imaging from ceiling or rear wall reflections.
Yes those are all good possiblities. With the exception of distortion in the recording chain imo, I mean it is possible but it seems a rookie mistake to have improper gain staging somewhere or faulty gear. Just guessing here, another possibility that recordings you disliked had violins either close-miced with clipon mics, or with an invisibly attached piezo mic. Those techniques can be excellent, but are also easier to sound harsh, like the use of SDC's.

Instruments have an optimal mic place (as determined by their players, who do recognise the best possible sound of their instrument). This seems to be contrary to my point in my post above about "artist intent", but that's because this optimal point is about 25-75cm from the instrument at a particular place, not somewhere in the hall.

I am not sure how you meant the part about Ambisonic mics letting you hear strings from the front? Because the 'front' part mic will also capture the other instruments than the strings and you can't separate them, and the 'rear mics' will also pick up the strings. Seems interesting.

Edit: I now understand your point about individualised HRTF's in combination with ambisonic mics. Yes that would be promising.
 
Yes those are all good possiblities. With the exception of distortion in the recording chain imo, I mean it is possible but it seems a rookie mistake to have improper gain staging somewhere or faulty gear. Just guessing here, another possibility that recordings you disliked had violins either close-miced with clipon mics, or with an invisibly attached piezo mic. Those techniques can be excellent, but are also easier to sound harsh, like the use of SDC's.

Instruments have an optimal mic place (as determined by their players, who do recognise the best possible sound of their instrument). This seems to be contrary to my point in my post above about "artist intent", but that's because this optimal point is about 25-75cm from the instrument at a particular place, not somewhere in the hall.

I am not sure how you meant the part about Ambisonic mics letting you hear strings from the front? Because the 'front' part mic will also capture the other instruments than the strings and you can't separate them, and the 'rear mics' will also pick up the strings. Seems interesting.
I suppose one thing is that I have yet to outfit my living room with a bunch of GIK Monster panels along with using those to improve my HRTF measurements. If you hadn't seen, my latest foray into binaural head-tracking is documented in https://www.head-fi.org/threads/rec...-virtualization.890719/page-121#post-18027627 (post #1,812). Sometimes I get a "you are there" with a recording, then something psychological or physiological changes and I am discontent again. For ambisonics, my hope is to capture the sound field and ITD around where an individual's ears would be at some ideal centered position in the hall, say, the fifth row (maybe this is an incorrect idealization of how ambisonic recordings are done). Buy out the adjacent seats if you have to to give the equipment enough room. With the ambisonic recording, each individual with good HRTF measurements and headphone compensation EQs should be able to transport their ears into that live hall. About "hearing strings from the front", I was describing instrument directivity information lost in stereo mixes. In the hall, I personally rarely perceive lateral reflections except with the treble content of some brass, reed, or percussion instruments on occasion, those types of instruments also being the most likely for me to perceive ceiling and front wall (I may have erroneously said "rear wall" earlier; the wall that you are facing) reflections; maybe there are even cases where due to plexiglass sheets and the filtering of the orchestra musicians themselves, I only really hear the ceiling reflection and not the direct sound. If I in Reaper with SPARTA AmbiRoomSim, ReaDelay, and ReaEQ try to upmix channels for simulating the far lateral and ceiling reflections, I might be able to get brass and whatnot to image from the "ceiling", but the strings will also erroneously image from such and dirty the sound. There is also the matter of how to capture what I perceive as being like almost 2D imaging of notes from the soundboard's reflection off of a concert grand's lid.
 
That is a very impressive piece of experimentation good done there, good writeup too. I have done very lightweight version of what you did, making a room emulation with EQ'd crosstalk and reverbs. I'm really interested in your work but I have to find a free moment to properly read that, possibly this Monday as I have free.

I don't have a headfi account. Considering your post didn't get much traction there, could you perhaps copy that post into a new thread here so that we can talk further? Edit: no problem if not of course.
 
Last edited:
I read it all including some of your other posts, impressive work. A man on a mission for audio nirvana :)

One question: what are you using as sources to test with? I'd assume an full surround or ambisonic recording, but you mentioned Blomstedt' Brahms, which is only available as stereo or 5.1 (as far as I know). If so, am I correct in assuming that your goal is to recreate a virtual concert hall based on the 5.1 version?

Thinking out loud, I think you then can't have a one-size-fits all virtual speaker placement or 5.1-to-binaural upscaler (don't know if these are the correct terms). Rather every recording requires you to reverse engineer what the mixing engineer did, in order to create the virtual concert hall. Unless you have pics, you'd need to do some assumptions (where was the orchestra, where were the microphones, and then what microphone ended up where in the mix). And then replace every track in a hall simulation, probably with some processing on there.
So basically remixing the recording. Difficult, but not impossible. I think...

If the above sounds like nonsense to you, keep in mind I have absolutely zero experience with binaural upscalers and I'm working on assumptions of what they do ;)
 
I read it all including some of your other posts, impressive work. A man on a mission for audio nirvana :)

One question: what are you using as sources to test with? I'd assume an full surround or ambisonic recording, but you mentioned Blomstedt' Brahms, which is only available as stereo or 5.1 (as far as I know). If so, am I correct in assuming that your goal is to recreate a virtual concert hall based on the 5.1 version?

Thinking out loud, I think you then can't have a one-size-fits all virtual speaker placement or 5.1-to-binaural upscaler (don't know if these are the correct terms). Rather every recording requires you to reverse engineer what the mixing engineer did, in order to create the virtual concert hall. Unless you have pics, you'd need to do some assumptions (where was the orchestra, where were the microphones, and then what microphone ended up where in the mix). And then replace every track in a hall simulation, probably with some processing on there.
So basically remixing the recording. Difficult, but not impossible. I think...

If the above sounds like nonsense to you, keep in mind I have absolutely zero experience with binaural upscalers and I'm working on assumptions of what they do ;)
Right now, I am only working with "perfecting" binaural stereo playback, mainly with recordings found on Idagio (I recently started a Qobuz trial and while being good for Hiromi has been mostly useless, but my credit card has already paid for an annual subscription; maybe once I implement "high-res" binaural head-tracking, I could listen to some recordings through Qobuz for the sake of listening to the "original"). Per my descriptions, stereo can be quite good for orchestra in most cases (except when string sections aren't imaged properly) and can sometimes seem to encode (whether or not by an error in my HRTF measurement or rendering) height information or that sense of vertical scale, perhaps mainly by bass content though I swear some recordings even through my Genelecs have conveyed height to the string reverb. It would just be missing those circumstantial hall reflections that I described, plus specific grand piano imaging. I rarely hear anything to the sides or behind me, so I don't think 5.1 surround would do the job (mind I like the idea of stereo subwoofers in an anechoic chamber); I would need some height channels. Indeed, without access to the original tracks, I could only think of really advanced AI convincingly upmixing stereo recordings for the instrument directivity within certain halls.

The "ambisonic" part is just one implementation of stereo to binaural rendering that provides flexibility for upmixing attempts or listening to surround recordings. Hopefully, whatever I come up with would be ready to play back good classical ambisonic recordings once such become more available.

For my current system, the main problem is the variability of my hearing with the same recording and DSP sounding harsh or "lifeless" or "wrong" at one part of the day, then suddenly a whole bunch of recordings sounding agreeable to great at another time or day.

To avoid hijacking this thread or cross-posting, we can continue the discussion on https://www.audiosciencereview.com/...opy-speaker-sounds-to-headphones.23154/page-5.
 
Most comments so far are focused on the HP frequency response characteristics, measurements and PEQ. Nobody has mentioned yet that we do not all have Harman curve hearing sensitivity. The biggest issue in your chain may be your ears. I have tested my ears frequency characteristics and use PEQ to compensate for my ear's sensitivity, not necessary my headphones frequency response. Test your ears, you may be surprised.
 
Most comments so far are focused on the HP frequency response characteristics, measurements and PEQ. Nobody has mentioned yet that we do not all have Harman curve hearing sensitivity. The biggest issue in your chain may be your ears. I have tested my ears frequency characteristics and use PEQ to compensate for my ear's sensitivity, not necessary my headphones frequency response. Test your ears, you may be surprised.
What do you mean exactly by compensating for your ear’s sensitivity?

In my case, I know I have some hearing loss in the mids on my right ear. I have tried to compensate for it (separate L/R EQ), but it does not work well… I think my brain naturally “compensate” for the L/R imbalance.
I also have losses in the highs, beyond ~12-14 kHz. I can “compensate” for it to a degree (treble boost/shelf), but it mostly means that i don’t need to be picky on the FR of an HP in the highs… It’s not that I hear nothing there, but by being less “sensitive” to highs freq., the different HP treble responses don’t matter much to me.
 
What do you mean exactly by compensating for your ear’s sensitivity?

In my case, I know I have some hearing loss in the mids on my right ear. I have tried to compensate for it (separate L/R EQ), but it does not work well… I think my brain naturally “compensate” for the L/R imbalance.
I also have losses in the highs, beyond ~12-14 kHz. I can “compensate” for it to a degree (treble boost/shelf), but it mostly means that i don’t need to be picky on the FR of an HP in the highs… It’s not that I hear nothing there, but by being less “sensitive” to highs freq., the different HP treble responses don’t matter much to me.
I mean pretty much what you described. If you are older and have steep roll-off over certain frequency - not much that you can do. I have less sensitivity to treble range in general. Depending on the recording I adjust treble shelf quite often. Also, in general, HiFiMan headphones with the reputation of "bright" are not too bright for me. However, I have some loss under 200HZ and PEQ adjustments works very well for me there. With L/R imbalance for certain frequencies I have tried PEQ and had an impression of the instrument's location horizontal movement in the sound stage. Just illusion? May be. Again, my point - if you do not know your ears looking at the graphs does not tell you what you may experience.
 
Agreed. Even if you were to standardise the use of the most common Decca Tree method. Intersting read:

But even if you standardise the microphones used and the size of the Decca Tree, every orchestra has a different number of musicians, and a different distance between each, in different halls, so you'd never end up with a 'true' recoridng. Not to mention that most listeners don't float above the orchestra.

Then there also no consensus on what is best. There are numerous other techniques, including where a microphone is laid down flat on the floor, and it sounds surprisingly good.
An orchestra is always arranged in a very specific way. About 5% of the live audience will ever sit where they can perfectly perceived the 2-dimensional layering (the height differences are immaterial). It is important to me to hear that in the recording and layout of classical pieces. Plenty of recordings from DG or Decca or Sony and others get that right. Recordings that don't are throwaways.
Many famous conductors are known to be very demanding when it comes to recording discipline and quality.
 
To be very fair, is not like headphones measurements are at the same level than Klippel speakers measurements, i considered a headphone measurement very poorly





Where is the IMD, ressoannce and more measurements?

FR in headphones is worthless because of the variancy of ears, and that is mostly what you get from asr. To be very honest with my opinion, i find the headphone measurements a kind of waste of time, the only worth measurement is the THD


ASR is not better than Head-Fi in measurements to be honest, in both sites you get fr and thd and done-...
 
Last edited:
To be very fair, is not like headphones measurements are at the same level than Klippel speakers measurements, i considered a headphone measurement very poorly





Where is the IMD, ressoannce and more measurements?

FR in headphones is worthless because of the variancy of ears, and that is mostly what you get from asr. To be very honest with my opinion, i find the headphone measurements a kind of waste of time, the only worth measurement is the THD


ASR is not better than Head-Fi in measurements to be honest, in both sites you get fr and thd and done-...
It is interesting that a number of others have the opposite opinion that FR measurements are the most important and that most THD results are barely audible anyway. I care about ease of EQing (smooth frequency response throughout) and not incurring too much distortion from raising any parts that needed raising; at least different sources of FR measurements can give one an idea of whether a headphone is close to neutral or by default "murders" the ear gain region. As I had documented in https://audiosciencereview.com/foru...phone-measurements.18451/page-61#post-1956895 (post #1,210), I am also interested in comparison of headphone IMD performance at least as seen through multi-tone FFTs, though this necessarily requires some form of EQ normalization for a fair comparison.
 
It is interesting that a number of others have the opposite opinion that FR measurements are the most important and that most THD results are barely audible anyway. I care about ease of EQing (smooth frequency response throughout) and not incurring too much distortion from raising any parts that needed raising; at least different sources of FR measurements can give one an idea of whether a headphone is close to neutral or by default "murders" the ear gain region. As I had documented in https://audiosciencereview.com/foru...phone-measurements.18451/page-61#post-1956895 (post #1,210), I am also interested in comparison of headphone IMD performance at least as seen through multi-tone FFTs, though this necessarily requires some form of EQ normalization for a fair comparison.
Your ''custom'' ear channel are going to change that anyway, its worthless
 
Your ''custom'' ear channel are going to change that anyway, its worthless
If the discussion was about frequency responce I would agree on this point. I not sure about distortion products and how it pertains to individual ear canal variation. It seems like a narrow topic, although I prefer lower THD whereas I appreciate the value of subjective evaluation of FR and distortion given the numerous unknown factors of headphone measurements.
 
Back
Top Bottom