Is EQ'ing headphones worth it?

[ernestcarl]

Do you use Android?

(probably my last) iPhone

 

[ernestcarl]

Last time I looked into and thought about loudness matching I think I concluded it was flawed. For a start you have to trust your ability to to actually be able to do it, and secondly the frequency response curve shouldn't be loudness matched across the whole frequency range right....I mean our ears aren't supposed to be sensitive to low bass frequencies and neither to high frequencies too, so with loudness matching wouldn't you just compensating for the deficits & injuries in your own hearing which is getting you further away from the "truth" of a real life experience, because real life auditory experiences are not frequency adjusted in loudness to match your personal hearing deficits....instead you adapt & live with those deficits? Or I'm thinking I'm missing some fundamental understanding of what loudness matching actually is?

It's not going to be an exact match, but one still might be able to achieve a close enough approximation.

IMO, you need to EQ each headphone close to neutral/flat before adjusting gain for each HP. A customized profile for each can (which can be loaded quickly) is preferable -- for a random assortment of HPs this is made easier if you have something like Sonarworks Reference for headphones or Morphit.

*You'd also probably want software that can play into two different DACs (with different DSP and volume settings) in parallel -- JRiver can do this type of operation.

 

[Robbo99999]

You might be missing a fundamental understanding of what neutral sound from headphones is. Organs in our ears, especially the pinna, make it so a neutral response isn't actually flat in the case of headphones. While I'm fuzzy on the subject (just know enough to pimp my cans well), this is due to various reflections, especially in the treble range, not being heard from headphones the same they would be heard from flat speakers.
You can get a better understanding of this by reading studies from Floyd Toole and Sean Olive.
So, with that said, the Harman Target and targets utilizing similar scientific context attempt to aim for neutral sound in accordance to popular preference, alongside measurements. Since not all ears are created equal, that's where loudness matching comes in.
Let's round it off with an example. Hypothetically, say a person has perfect hearing. In that case, loudness matching is not compensating for hearing deficiencies, rather, it is compensating for the shape and organ placement of their ear. Reference sound from headphones to the person with perfect hearing and a typically formed ear still wouldn't be flat.
Now, for the person with comparable hearing and an atypically formed ear, that's where loudness matching may have an advantage in certain respects over things like the Harman Target.
However, loudness matching in compensation for hearing deficiencies is also viable. You say it would get you further from the truth- nay, the hearing deficiencies themselves do.
With hearing deficiencies, you aren't exactly hearing the truth in "real life" either.

I'm somewhat familiar with pinna and also understand how the Harman Target is created for both speakers & headphones. I either didn't explain myself clearly or you missed my point, but it's a combination of both. I think it's folly to match equal loudness to each frequency point in the hearing range if that's what loudness matching entails. This is because our natural hearing sensitivity varies accross the frequency range, therefore if you loudness match each frequency to the same perceived loudness across the frequency range then you're gonna be creating an unnatural curve, I just can't see how it's valid....afterall the sounds that occur in the environment aren't artificially loudness matched in their frequency profile to every single individual listeners hearing curve - instead our ears have to receive whatever sounds are produced in the environment and then we interpret those sounds. If you loudness match every frequency on the curve to be the same perceptible volume then it is gonna be nothing like a natural sound because you'd be boosting & recessing frequencies that you're insensitive or sensitive to, thereby creating unnatural conditions.

The truest theoretical way you could create a headphone target curve for yourself, that would display for that individual a realistic sound profile, would be the following procedure/steps:

1. place measuring microphones directly on your eardrum, and then play sinesweeps on perfectly flat speakers in an 'ideal room' and record the frequency response of those microphones on your eardrum. That frequency response would be your own personal target curve for natural sound that takes into account both your HRTF (shape of your head), as well as the shape of your ear and ear canal into account. (For various reasons due to the way bass 'feels' on headphones you'd then probably have to put in a bit of a bass boost to preference....part of the Harman research).
2. So in step 1 above you've got your own personal target curve for natural sound at your eardrum, now the next part is to calibrate your headphones so that produce an identical curve across the frequency range when measured at the eardrum. So you'd put your headphones on your head & run sine sweeps and measure the output from the mics located on your eardrum, and then you'd create an EQ profile to manipulate the headphones frequency response to match your target curve created in Step 1. Then job done!
3. But....but....that's not possible to do so both Step 1 & Step 2 have to be approximated. Step 1 is approximated from the Olive & Toole research where they created the Harman Curve for headphones from 'analysing/testing lots of subjects' in those conditions....I don't know the full details of how they did this. And Step 2 is approximated by having a dummy head and ears on which the headphone is measured, for the purpose of EQ'ing the headphone to the target curve (Oratory1990 having the best measuring rig & expertise).

I think there's fundamental flaws with using loudness matching accross the frequency range as a means to EQ a headphone for the simple reason that our heads & ears aren't 'designed' to hear all frequencies at the same volume/loudness...that's the fundamental point, therefore loudness matching across the whole frequency range to the same perceived volume is totally flawed. That is if I understand the loudness matching process, or have I missed something?...that's not sarcasm by the way, because I admit I've not read up on loudness matching but gained snippets from here & there.

 

[Dreyfus]

My take on that ...

Calibration infront of a real speaker:
- subjectively flat
- renders the source infront of the listener
- includes full HRTF
- can be recorded in DF (room sound) or FF (no room sound)

Calibration with noise only:
- subjectively flat
- renders the source left and right of the head
- includes ear resonances and loudness contour
- recorded in FF (no room sound)

IMO there is no right or wrong, only different standards and tastes.

Music that has been produced towards studio standards (infront of loudspeakers) may sound very convincing with an impulse response emulating your HRTF infront of a flat speaker. Productions made with varying speaker and room responses may at least sound okay.

On the other hand, the calibration to noise only (without frontal localization) may sound better with binaural recordings or mixes which have gone through headphones. Furthermore, it is far better than just listening with a strangely colored headphones (take the "Beyerdynamic Peak" as an example), grants the maximum hearing sensitivity across the whole spectrum (great for gaming!) and ... it is practically cost-free to create.

By the way, the noise only calibration can also be a starting point for the HRTF loudspeaker calibration:

Regards
P48

 

[Robbo99999]

My take on that ...

Calibration infront of a real speaker:
- subjectively flat
- renders the source infront of the listener
- includes full HRTF
- can be recorded in DF (room sound) or FF (no room sound)

Calibration with noise only:
- subjectively flat
- renders the source left and right of the head
- includes ear resonances and loudness contour
- recorded in FF (no room sound)

IMO there is no right or wrong, only different standards and tastes.

Music that has been produced towards studio standards (infront of loudspeakers) may sound very convincing with an impulse response emulating your HRTF infront of a flat speaker. Productions made with varying speaker and room responses may at least sound okay.

On the other hand, the calibration to noise only (without frontal localization) may sound better with binaural recordings or mixes which have gone through headphones. Furthermore, it is far better than just listening with a strangely colored headphones (take the "Beyerdynamic Peak" as an example), grants the maximum hearing sensitivity across the whole spectrum (great for gaming!) and ... it is practically cost-free to create.

By the way, the noise only calibration can also be a starting point for the HRTF loudspeaker calibration:

Regards
P48

I watched all of that video, it's a good way of doing loudness matching so I give it that, but none of that addressed my fundamental concerns regarding the validity of loudness matching....in fact all of my prior assertions about loudness matching in my previous post remain - ie that due to our heads (HRTF) & due to 'everything to do with' our ears we aren't 'designed' to hear all frequencies at the same volume/loudness...that's the fundamental point, therefore loudness matching across the whole frequency range to "force" to the same perceived volume is totally flawed....we're not supposed to be able to hear content across the whole frequency range at the same volume. I think it's a totally fundamentally flawed process. (Unless I'm still missing a step in the process/understanding, but I don't think I am)

I think the only way to go about accurate EQ is through physical measurements. Olive & Toole did it for speakers by equating spins (frequency measurements) of speakers with test subject speaker preference which resulted in their Harman Curve for loud speakers in a room, so 'we' EQ to that measured target curve, and more times than not this results in a better subjective experience. They also did the same things for headphones, which I described the rough process in my previous post. I think the only valid fine tuned accurate way of going about EQ is based on physical measurements of speakers and headphones, and adjusting accordingly to the target curves that have been created by the research....and yes you can change those target curves a little because they (Olive, Toole) state themselves that broad tone control (bass/treble) are needed in everyday listening to account for differences in recording of the source and to account for personal preference (which is mainly focussed around how much bass boost or lack thereof). So, yes, these are approximations & averages like I explained in my previous post in more detail, but I am so far still convinced that loudness matching is a fundamentally flawed process that would give unnatural results due to the reasons I explained in this post & my previous.

 

[solderdude]

Consider most recordings are mixed at around 75 to 80dB average SPL so there you have your reference. You Just have to EQ at that level. If so desired add bass to taste or level. There really is nothing more to it. You are overthinking this.
Note that while you bombarded the K702 as reference (have gotten used to it) it really isn't a reference at any SPL.

 

[Robbo99999]

Consider most recordings are mixed at around 75 to 80dB average SPL so there you have your reference. You Just have to EQ at that level. If so desired add bass to taste or level. There really is nothing more to it. You are overthinking this.

I'm not overthinking it until you prove me otherwise. Are my points not valid or do you not understand them. I would think you understand my points, if you do then it would be about educating me & letting me see the light....I'm open to a different/better understanding. I do think I have an understanding & visualisation of how frequency curves/response fit into what would theoretically be measured at the eardrum (HRTF, ear canal effects on frequency). From the way I understand how it all fits together then loudness matching is fundamentally flawed.

(K702 is irrelevant to the discussion, I'm beginning to question your 'abilities')

 

[solderdude]

It isn't flawed at all. When you have flat speakers playing in anechoic conditions you van EQ heaphones similar. You van also EQ to a known target curve. Using a HATS is fundamentally wrong though IMO.

 

[Robbo99999]

It isn't flawed at all. When you have flat speakers playing in anechoic conditions you van EQ heaphones similar. You van also EQ to a known target curve. Using a HATS is fundamentally wrong though IMO.

I don't fully picture what you're saying, but I think you're talking about something slightly different. Are you saying that if you personally have flat calibrated speakers that are playing in an anechoic room then it's possible to take some headphones, put them on your head and then subjectively EQ them to sound like the speakers by taking the headphones on & off your head whilst alternately listening to the speakers....by playing the same test tone on the speaker and on the headphone? (That's a lot of conditions you have to have available to you.) So that would be loudness matching between a speaker and headphone, that's also not the same as loudness matching we're talking about in this thread. In this thread we're talking about loudness matching frequencies psycho acoustically across the whole frequency range so all frequencies have the same loudness, and this would be done on one set of headphones (no loudness matching between 2 sets of speakers or headphones) as a means of supposedly EQ'ing a headphone 'correctly to a natural/real sound' - in my understanding that's a fundamentally flawed process for the reasons I explained in my previous posts.

And your second point about "can also EQ to a known target curve". How can you EQ by subjective listening to a known target curve if you've got nothing to reference against. So I don't understand that point you made.


EDIT: and here's a post from someone else that has the same understanding as me re Equal Loudness EQ being fundamentally flawed, so figured I'd post it here as another way of explaining what I've been getting at in my previous posts:
https://hydrogenaud.io/index.php?PH...0vjv2cmukod2&topic=111845.msg921737#msg921737
He says the following:
--------------------------------------------------------------------------------------------------------------------------
I'm not sure why people that discover the equal loudness contours it are so attracted to it, but I think their reasoning goes like this:
A headphone's frequency response has peaks and dips, therefore these frequencies will be played too high or low in volume, therefore correcting for equal loudness contours fixes all issues.

But that is simply wrong. A sound source that produces a flat frequency response will result in everything but a flat frequency response at your eardrum. For example, your ear canal can easily do a +10 dB boost between 2-3 kHz and this is perfectly natural and desired. Your pinna, head and even torse will have a significant effect too. Additionally, our hearing is also most sensitive in that range so if you compare sine tones it may very well sound 15 dB louder, but that's supposed to be that way.

If you attenuated that range by 20 dB then you could easily be 15 dB off.
------------------------------------------------------------------------------------------------------------------------------

 

[ernestcarl]

... correcting for equal loudness contours fixes all issues.
--

Not sure what the big deal is all about as I really don’t think anyone is trying to match that, but...

I already use some form of loudness compensation in all my drivers, full-sized headphones and speakers. I also know I can live without it. Though I am really not aware of anyone who explicitly claims “equal loudness contours fixes all issues”. So I’m not exactly certain who you are arguing strongly against here...

“Loudness matching”, as I understand it, between several different drivers and at different volume levels is an exercise in approximation. And I think you already agree to that...

Just because it’s an imperfect approximation does not necessarily mean the outcome should always sound “unnatural” to the listener — and, honestly, why bother all these years with such unnecessary and “unnatural” DSP if I felt that it did more harm than good?

I regularly listen -30dB below my reference — occasionally -40dB at night — as far as my experience goes, my eardrums have not suffered unduly due to an auditory sensation of severe synthetic unnaturalness.

 

[Robbo99999]

Not sure what the big deal is all about as I really don’t think anyone is trying to match that, but...

I already use some form of loudness compensation in all my drivers, full-sized headphones and speakers. I also know I can live without it. Though I am really not aware of anyone who explicitly claims “equal loudness contours fixes all issues”. So I’m not exactly certain who you are arguing strongly against here...

“Loudness matching”, as I understand it, between several different drivers and at different volume levels is an exercise in approximation. And I think you already agree to that...

Just because it’s an imperfect approximation does not necessarily mean the outcome should always sound “unnatural” to the listener — and, honestly, why bother all these years with such unnecessary and “unnatural” DSP if I felt that it did more harm than good?

I regularly listen -30dB below my reference — occasionally -40dB at night — as far as my experience goes, my eardrums have not suffered unduly due to an auditory sensation of severe synthetic unnaturalness.

Hmm, I don't think you understand the fundamentals I was getting at, and I can't really keep typing them out. I've explained as best as I can in my previous posts why I understand Equal Loudness EQ tuning accross the frequency range to be fundamentally flawed and unnatural. I don't really know how many different ways I can explain the mechanisms for why that is. I think if a person understands & can visualise what is happening in how the Harman Curve was created for both speakers & headphones then I think you'd have to conclude that Equal Loudness Matching as an EQ concept is fundamentally flawed and completely at odds with what has been learned & proven by the Harman Research. Unless I'm missing a fundamental step in my understanding of how it all fits together then I'm not wrong. The problem is that anyone that has replied to me actually hasn't quoted the parts of my post where I describe/visualise the problems with Equal Loudness EQ matching, so the discussion can't be moved forward.

To be clear I'm not talking about using Loudness Matching to match 2 different headphones or sets of speakers (one of which might be calibrated to a known standard).....I'm talking about using Loudness Matching across the frequency range just on (within) one set of headphones or speakers....as the guy in the video a few posts above me is doing.

(go back and read & try to understand my string of posts starting from this one yesterday if you've not been following the whole thing: https://www.audiosciencereview.com/...s-eqing-headphones-worth-it.11523/post-456047)

 

[Robbo99999]

Although equal loudness curves might be a little bit of part of the puzzle I'm not taking into account involved in my prior visualisations re Equal Loudness EQ's being invalid when compared to say the Harman Curve as there are indeed some inverse relationships going on with the shapes when compared to Harman Curve:
https://en.wikipedia.org/wiki/Equal-loudness_contour#/media/File:Lindos1.svg

But I don't think the above describes the whole story/picture, and furthermore you can see that Equal Loudness has it's own set of pitfalls/limitations associated with it being volume level dependant - so I still think Equal Loundess EQ is fundamentally flawed and at very best may only be a small part of the correct overall real picture.

So my last thoughts in terms of doubting my own prior assertions in my previous posts re Equal Loudness EQ tuning to be fundamentally flawed would be based around psycho-acoustically how loud to we perceive the frequencies that actually hit our ear drum (the frequencies that would be defined by the shape of the headphone Harman Curve for instance), note that's not the same thing as the loudness curves picture above in this post, because those loudness curves above are from loud speakers and don't describe attenuation of frequencies by HRTF and ear canal. So the Harman Headphone Curve might describe the actual frequencies that make it past our head down our ear canal, but do we really perceive the loudness of those frequencies linearly to the dB level that they appear at the eardrum. So there are some unanswered parts of my thinking that could be gaps in my assertions that Equal Loudness EQ tuning is fundamentally flawed. But I'm playing devils advocate here because no one is addressing the specifics of my understanding to either prove or disprove them. As it currently stands I do see Equal Loudness EQ to be be fundamentally flawed still and at the very least not the best way of arriving at true natural sound.

EDIT: and this is not even accounting for the fact that Equal Loudness EQ will be compensating for an individuals personal hearing deficiencies that they either grew up with or developed during their lifetime....so unless you believe a headphone experience should be a hearing aid experience then you're not getting a true representation.....unless you actually wear hearing aids in which case you'd take out your hearing aids when listening to your music and then the equal loudness EQ would go some way to replacing the action of your hearing aids. It's a bit of a toss up, but equal loudness tuning would be de-tuning your normal hearing curve you grew up with (to some degree).

 

[ernestcarl]

Hmm, I don't think you understand the fundamentals I was getting at, and I can't really keep typing them out. I've explained as best as I can in my previous posts why I understand Equal Loudness EQ tuning accross the frequency range to be fundamentally flawed and unnatural. I don't really know how many different ways I can explain the mechanisms for why that is. I think if a person understands & can visualise what is happening in how the Harman Curve was created for both speakers & headphones then I think you'd have to conclude that Equal Loudness Matching as an EQ concept is fundamentally flawed and completely at odds with what has been learned & proven by the Harman Research. Unless I'm missing a fundamental step in my understanding of how it all fits together then I'm not wrong. The problem is that anyone that has replied to me actually hasn't quoted the parts of my post where I describe/visualise the problems with Equal Loudness EQ matching, so the discussion can't be moved forward.

To be clear I'm not talking about using Loudness Matching to match 2 different headphones or sets of speakers (one of which might be calibrated to a known standard).....I'm talking about using Loudness Matching across the frequency range just on (within) one set of headphones or speakers....as the guy in the video a few posts above me is doing.

(go back and read & try to understand my string of posts starting from this one yesterday if you've not been following the whole thing: https://www.audiosciencereview.com/...s-eqing-headphones-worth-it.11523/post-456047)

Watched the video, though, obviously, that is not what I had specifically in mind when the phrase “loudness matching” came about.

 

[ernestcarl]

Just going to get a cup of coffee for as it’s still still early in the morning here. And will re-attempt to digest again your earlier posts and the main point you're repeatedly driving at...

 

[Robbo99999]

Watched the video, though, obviously, that is not what I had specifically in mind when the phrase “loudness matching” came about.

Yep, I thought so when I read your post.

Just going to get a cup of coffee for as it’s still still early in the morning here. And will re-attempt to digest again your earlier posts and the main point you're repeatedly driving at...

Ha, yeah!

 

[Robbo99999]

Final summation on thoughts on ways to describe Loudness Matching EQ process to assess if it's a valid process.

In order for Loudness matching EQ to be valid then the following needs to be true when listening to speakers in a room:

1. We hear all frequencies across the spectrum from 20Hz to 20kHz equally in terms of perceived volume when the speaker produces the tone at the same dB.
2. All frequencies across the spectrum from 20Hz to 20kHz should be perceived volume matched for a natural sound in a room.
3. It's OK for peculiarities in a normal individuals hearing curve (that they've been born with or grown up with and/or developed due to injuries) to be overridden by loudness matching each frequency on curve.

And answering my own points from above in the same order:

1. We don't perceive frequencies across the spectrum from 20Hz to 20kHz equally in terms of perceived volume as seen by the loudness contour graph.
2. A natural sound in a room shows decay of higher frequencies to a greater extent than the lower frequencies, and so this is not accounted for in perceived volume matching across the frequency range (at least not reliably so as seen by the shape of the equal loudness curves).
3. I think changing your natural 'deficiencies' hearing curve that you've grown up & lived with and/or covering up hearing deficiencies by loudness matching is an unnatural experience probably too far away from your everyday auditory experience....so I think this would make music sound 'ungenuine'/unintended/weird. (You'd be going against/changing your own "psycho-acoustic imprint" that you have learned & adapted to over your entire life during all your waking hours)

Yeah, I don't see Loudness Matching EQ technique as valid.

 

[solderdude]

You can NOT EQ to equal loudness contours. Nor does one need to take it into account either. Recordings are mastered at around 80dB average. That is the only reference there is. That reference is flat speakers and sound engineer based.
It also had nothing to do with eardrums freq. Response either.
The described methode works and non HATS measurements work. Then you van add bass acc. To Harman target or EQ to other targets.

It really is that simple

 

[Robbo99999]

You can NOT EQ to equal loudness contouren. Nor does one need to taken it into account either. Recordings are mastered at around 80dB average. That is the only reference there is. That reference is flat speakers and sound engineer based.
It also had nothing to do with eardrums freq. Response either.
The described methode works and non HATS measurements work. Then you van add bass acc. To Harman target or EQ to other targets.

It really is that simple

Wow, really, so to be clear you agree with that video that was posted earlier by a member, just so I'm clear on what you mean:

You agree with the validity of that video as a valid method?

 

[solderdude]

Yes provided you use noise bands, his program, and do the matching at around 75 to 80dB.
The dangerous part is you include your own HRTF and need to be familiar with such methods of testing.
Griesinger of course knows what he is doing. Some clueless person will f this up for sure.

 

[Robbo99999]

Yes provided you use noise bands, his program, and do the matching at around 75 to 80dB.
The dangerous part is you include your own HRTF and need to be familiar with such methods of testing

Right, I'm surprised to hear you agree with the concept of Equal Loudness EQ. As I've been trying to show it's flawed on many levels, I see practically zero validity in it. You haven't really technically described to me why you think it's accurate nor have you tried to refute my arguments as to why it's not valid which I think are very clear & obvious, especially in this post (https://www.audiosciencereview.com/...s-eqing-headphones-worth-it.11523/post-456884), so you either can't be bothered or you don't really understand it....I would think you understand or at least I'm sure you think you understand it because you're heavily involved in this industry so to speak, but I'm not seeing it. I've been open to being persuaded otherwise, but no one has countered the specifics of my points or said....hang on Rob you're missing this point here, or actually you've not taken this into account, etc.