Headphones and the Harman target curve

[bluefuzz]

I'm afraid that's a tad wide of the mark.

In what way? Obviously it’s a simplification, but the point is that ultimately the only way to corrolate visual or auditory experience to the real physical world is to ask people what they are seeing or hearing because we can’t directly measure that subjective sensory experience.

The point is that any broadcasts or discs you see are mastered to that standard.

Yes, but that’s rather a different discussion.

Now, if all audio mixing and mastering studios had speakers (& rooms) and headphones all EQd to the Harman Curve, you'd have a point

Not sure what it has to do with my point. In an ideal world, perhaps it would be nice if music were mixed and mastered to some agreed standard, although I’m not sure it would help for anything other than live recording of acoustic instuments. Anyway, that horse has long since bolted.

surely it must be to hear from our headphones exactly (or as near as possible) what the engineer heard when creating the master.

Most music is mixed and mastered to sound acceptable on a wide range of equipment from cell phones, car audio systems to smart speakers and maybe even high end hifi. There will never be 'one true' correct listening device or standard. As I’ve probably said before I think chasing ‘what the producer heard in the studio’ is a fool’s quest and an impossible chimera. All we can hope to achieve is fidelity to the final recording as shipped on cd, internet stream or vinyl ...

 

[Yorkshire Mouth]

Yes saying measuring listener preferences is not scientific makes no sense, and no arguments are made to support it.

I have actually said this several times, but for yourself I don't mind repeating.

Yes, you absolutely can measure preference scientifically, and that's clearly what you've done here. I have already said this, and I'm in no way critical of that. Indeed, I'll go as far as to say you're almost certainly spot on with your conclusions, given the extensive testing you've done, and the methods used. The Harman Target Curve is almost certainly an accurate summation of an average preference of the people involved in your tests, and that's likely to be replicated across the general public.

Sometimes, given the limitations of internet forums, we use a certain amount of shorthand.

My comments are not in any way a criticism of your work. They're a commentary on the continued use of your scientific measurement of preference to try and portray it as meaning that headphones which follow the Harman Target Curve are scientifically proven to be a more accurate reproduction of the source than headphone which do not.

Too many people are saying "The Harman Target Curve is right", and them continuing to discuss that as meaning the curve is a more accurate reproduction of the source, than accurate to an average preference. I think we'll agree that two aren't necessarily the same thing.

If anyone says a headphone is 'wrong' or 'bad' solely because they don't follow the curve, I feel that's a misrepresentation of your work, though I'll leave you to comment on that. Sadly, it's something that many people regularly do.

 

[Yorkshire Mouth]

In what way? Obviously it’s a simplification, but the point is that ultimately the only way to corrolate visual or auditory experience to the real physical world is to ask people what they are seeing or hearing because we can’t directly measure that subjective sensory experience.



Yes, but that’s rather a different discussion.



Not sure what it has to do with my point. In an ideal world, perhaps it would be nice if music were mixed and mastered to some agreed standard, although I’m not sure it would help for anything other than live recording of acoustic instuments. Anyway, that horse has long since bolted.


Most music is mixed and mastered to sound acceptable on a wide range of equipment from cell phones, car audio systems to smart speakers and maybe even high end hifi. There will never be 'one true' correct listening device or standard. As I’ve probably said before I think chasing ‘what the producer heard in the studio’ is a fool’s quest and an impossible chimera. All we can hope to achieve is fidelity to the final recording as shipped on cd, internet stream or vinyl ...

I have to be honest, reading my comments and your comments, I don't think we're much in disagreement.

The one thing on which I'll differ is that I don't think it would be impossible to create a standard. If we decided on a particular curve, and asked studios to test to see if their speakers/mixing rooms and/or headphones used (both possibly with EQ) conformed to that, I believe it would be possible.

Indeed, I'll go further. I think Harman's 'great room' is proof that it can be done. And Amr has shown here that there are headphones which, if EQd, can be made to follow the Harman curve.

 

[bluefuzz]

don't think it would be impossible to create a standard

I would think most decent studios already conform to the standard of anechoically flat monitors (Genelecs, Neumans etc.) in a deadish room. But that’s not the problem. The problem is music has to be released into the wild and played on all sorts of more or less accurate devices in all sorts of less than salubrious environments. In the case of headphones all we can do is ask people how they want that music to sound and we end up back with the Harman curve ...

 

[Yorkshire Mouth]

I would think most decent studios already conform to the standard of anechoically flat monitors (Genelecs, Neumans etc.) in a deadish room. But that’s not the problem. The problem is music has to be released into the wild and played on all sorts of more or less accurate devices in all sorts of less than salubrious environments. In the case of headphones all we can do is ask people how they want that music to sound and we end up back with the Harman curve ...

Headphones remove the room from the equation.

It is possible to measure the frequency response in the mixing studio (as you’ve mentioned) then tune headphones to that.

Indeed, to an extent the Harman research did something pretty similar. But it then played that back to listeners and gave them treble and bass controls.

 

[MayaTlab]

Indeed, to an extent the Harman research did something pretty similar. But it then played that back to listeners and gave them treble and bass controls.

I may be mistaken but my understanding was that the loudspeakers in room response was equalised to flat, the headphones equalised to the resulting response on a dummy head, and then listeners were given bass and trebles controls for both, and happened to prefer a resulting curve that matched reasonably well.

 

[preload]

Well, it's my job to arrange controlled Covid vaccination clinics in the UK for the NHS, for which I've had to plough through all of the literature on the initial trials, and decisions on whether or not to licence for use by the MHRA . And you're wrong.

If you're involved in the UK vaccine effort, and you claim to have read the vaccine trials, and you did NOT see that the vaccine RCT's routinely provided patient-reported symptoms in their Results, then that leaves me sad and speechless.

Well, I've answered that already several times. If you can't be bothered to read the posts I've already contributed, I don't think one more will make a difference.

So no, tough. Go and read what I've said already.

Nah, I didn't think you actually had a point here, but wanted to confirm. Confirmed.

 

[Yorkshire Mouth]

I may be mistaken but my understanding was that the loudspeakers in room response was equalised to flat, the headphones equalised to the resulting response on a dummy head, and then listeners were given bass and trebles controls for both, and happened to prefer a resulting curve that matched reasonably well.

AFAIA you're correct. I believe the listeners just boosted bass by 4dB, or so (as an average).

If you look at the Harman curve, it goes up by around 4dB working backwards from 100hz to 20hz. That rise is the pretty much the amount the bass was turned up (on average). I believe that without it, the curve was pretty flat in that area, though I stand to be corrected.

 

[Feelas]

I'm not saying the Harman curve is wrong. For all sorts of reasons I suspect it's right, or close. I'm just saying you can't say it's 'scientifically correct' in anything other than a measurement of preference, as opposed to accuracy.

To be fair, I'm not sure it's supposed to be anything else, but appears to have become so in some people's minds.

It is not supposed to be anything else, but strangely (well... not really, since subjects were asked to reproduce the room experience using the EQ) the perceived FR seems to correlate with known preferred room-curves for speakers, and as such it can be treated as an accurate target for reproduction of recorded music, since nobody targets unnaturally (read: studio-like) treated rooms for resultant music.

AFAIA you're correct. I believe the listeners just boosted bass by 4dB, or so (as an average).

If you look at the Harman curve, it goes up by around 4dB working backwards from 100hz to 20hz. That rise is the pretty much the amount the bass was turned up (on average). I believe that without it, the curve was pretty flat in that area, though I stand to be corrected.

The bass boost is to be expected in typical untreated room conditions, sometimes even higher than 4dB. No surprise there.

No need to mince words or have any confusion. It's a scientifically accurate measurement of FR preference in a population in the conditions in which they were performed.

FWIW I remember that Floyd Toole reported that there is no real variation between populations from various world regions, which you'd assume should happen, since house construction is different and thus speaker-preference should be affected by what works better in specific housing types. I think that gives us potential to treat the results as generalized. Some other research has shown that people generally seem to rank the most accurate speakers the best, which seems resonable to also assume to hold w/ other types of audio reproduction equipment.
Please discuss that bit in the future, since I guess that idea that people would generally target accurate gear in blind conditions seems not to go around the heads in here. Seems that some in the community still dwell on regarding the "general public as being bass-friendly and we don't do that here, we like bright ear-piercing treble", a mundane effort in tribalism.

That might actually be why the study is considered to be pretty accurate on what is preferred by the general population. Quite unsurprisingly there was some joint work done between Harman-curve buds & Toole, which can be found in various papers. Yet that isn't common knowledge if you don't even check the most basic facts.

 

[Yorkshire Mouth]

It is not supposed to be anything else, but strangely (well... not really, since subjects were asked to reproduce the room experience using the EQ) the perceived FR seems to correlate with known preferred room-curves for speakers, and as such it can be treated as an accurate target for reproduction of recorded music, since nobody targets unnaturally (read: studio-like) treated rooms for resultant music.


The bass boost is to be expected in typical untreated room conditions, sometimes even higher than 4dB. No surprise there.



FWIW I remember that Floyd Toole reported that there is no real variation between populations from various world regions, which you'd assume should happen, since house construction is different and thus speaker-preference should be affected by what works better in specific housing types. I think that gives us potential to treat the results as generalized. Some other research has shown that people generally seem to rank the most accurate speakers the best, which seems resonable to also assume to hold w/ other types of audio reproduction equipment.
Please discuss that bit in the future, since I guess that idea that people would generally target accurate gear in blind conditions seems not to go around the heads in here. Seems that some in the community still dwell on regarding the "general public as being bass-friendly and we don't do that here, we like bright ear-piercing treble", a mundane effort in tribalism.

That might actually be why the study is considered to be pretty accurate on what is preferred by the general population. Quite unsurprisingly there was some joint work done between Harman-curve buds & Toole, which can be found in various papers. Yet that isn't common knowledge if you don't even check the most basic facts.

You may well be right in some or all of that.

It doesn’t change the fact that some people are treating the curve as something other than a preference curve, when that’s not what its creators have claimed.

 

[preload]

You may well be right in some or all of that.

It doesn’t change the fact that some people are treating the curve as something other than a preference curve, when that’s not what its creators have claimed.

It seems to bother you a lot that people wish to "target" a curve because it sounds better (i.e. "preferred") to them. Most consumers and enthusiast simply want a stereo system that sounds good to them. Measurements and so-called "accuracy" are simply ways to help get there.

In other words, blinded listener preference is the ultimate goal for most people, not this ill-defined concept you have of "accuracy." The one exception might be recording/mastering professionals (wrong forum for that).

 

[Yorkshire Mouth]

It seems to bother you a lot that people wish to "target" a curve because it sounds better (i.e. "preferred") to them. Most consumers and enthusiast simply want a stereo system that sounds good to them. Measurements and so-called "accuracy" are simply ways to help get there.

Not at all. I’m only bothered by people who use that curve as a gold standard, with anything not confirming to it being empirically ‘wrong’.

I don’t believe chasing accuracy at Audio Science Review should ever be sneered at, how ever difficult it is to define.

 

[ADU]

You may well be right in some or all of that.

It doesn’t change the fact that some people are treating the curve as something other than a preference curve, when that’s not what its creators have claimed.

Fwiw, here is some of what Harman is claiming...

They're claiming that it is "very neutral and accurate" for a majority of listeners, which goes well beyond the notion of just a preference curve.

And it also seems fairly clear from some of Omid's other comments above that some of their objectives for their target headphone response curve were also to reduce the coloration in headphones to better approximate the response of neutral loudspeakers in a room. And if that is their stance, then I think it's fair game to discuss, debate and question the reliability and precision of their results in all of the above respects. Others may (and do seem) to disagree though.

If they had simply said, "it's just about people's preferences, and accuracy and neutrality wasn't really of our goal", then maybe you'd have a somewhat stronger case for your above argument, Yorkshire Mouth. If you look at the actual research though, much of it is was clearly centered around the measured responses of loudspeakers in a room. Because that's the type of sound that the subjects in their early tests clearly seemed to prefer.

And while they did find some fairly noteworthy discrepancies among certain groups of listeners, whether those later tests showed any meaningful variations among the general population is harder to establish. Because the system they were using for their in-ear tests of both headphones and loudspeakers wasn't designed to be that accurate in the frequency ranges where these variations might have occurred. And because the discrepancies in many of the studies were fairly small. So more work really needs to be done to further refine, and confirm or deny their previous results. Preferably with a somewhat better designed measurement rig.

I understand, btw, why Harman is interested in both accuracy and preference. Because there are consumers and markets for both types of products.

I personally don't care about the subjective preferences though. All I really want to know is-- how do several good pairs of neutral loudspeakers with a flat direct response, good bass extension, and a fairly (though not necessarily perfectly) linear off-axis response measure at the eardrum when they're put into a typical semi-reflective domestic listening space. And I'd love it if sites like ASR, Rtings, Head-Fi, Headphonesdotcom, Oratory, Crin, and the manufacturers of the rigs used by these folks (HBK, GRAS, Head Acoustics...), and maybe also Harman would all put their heads together, and try to figure that out a little better. So we'd have a better and more widely compatible and reproducible standard for this sort of thing.

Since this is clearly something that alot of other folks are also interested in, I don't really know why something like this hasn't already been done before. And why there's so much resistance to it.

Until such a time as something like that happens though, I guess the Harman curve is probably it (unless you're open to some other "whacky" altermatives, like the ones I've been discussing recently a little here).

 

[preload]

Not at all. I’m only bothered by people who use that curve as a gold standard, with anything not confirming to it being empirically ‘wrong’.

I also don't think Harman curves should be used as the gold standard, because they're not, and their publications are clear about that. The curves ARE perhaps the best model so far in predicting listener preferences, but they're not perfect. The other issue is that the curves are perhaps better represented as a shaded region (or at least error bars) instead of a thin line in order to represent the degree of variation, uncertainty, and to account for different conditions.

I don’t believe chasing accuracy at Audio Science Review should ever be sneered at, how ever difficult it is to define.

"Accuracy" is easy to define via the well understood suite of measurements for solid state devices, like DAC's and amplifiers - and is what pretty much everyone is looking for (with some exceptions). "Accuracy" is not easy to measure or quantify with transducers, whose measurements are represented by a 3-d sphere and have a time-domain, and the relationship between what is measured and what is heard is not completely understood. So by all means, continue "chasing accuracy" for transducers and see how far that goes.

 

[Yorkshire Mouth]

Hi Adu, that’s an excellent post, and I agree with almost all of it. I’ll post a bit more detail when I get to my PC.

In the meantime, and to lighten the mood a little, as we’re talking about sound quality, here’s the legend Bobby Whitlock and his opinions on the vinyl mix fit the recent re-release of George Harrison’s classic All Things Must Pass.

In many ways it’s quite a sad state of affairs, but it’s impossible to watch and listen to his response without laughing.

WARNING: Contains ‘industrial’ language.

 

[Jimbob54]

Hi Adu, that’s an excellent post, and I agree with almost all of it. I’ll post a bit more detail when I get to my PC.

In the meantime, and to lighten the mood a little, as we’re talking about sound quality, here’s the legend Bobby Whitlock and his opinions on the vinyl mix fit the recent re-release of George Harrison’s classic All Things Must Pass.

In many ways it’s quite a sad state of affairs, but it’s impossible to watch and listen to his response without laughing.

WARNING: Contains ‘industrial’ language.

His system must not be resolving enough.....

 

[Feelas]

I also don't think Harman curves should be used as the gold standard, because they're not, and their publications are clear about that. The curves ARE perhaps the best model so far in predicting listener preferences, but they're not perfect. The other issue is that the curves are perhaps better represented as a shaded region (or at least error bars) instead of a thin line in order to represent the degree of variation, uncertainty, and to account for different conditions.

Using Harman curves as a "gold standard" would help things, since anything else could be assessed in terms of deviance from a known common curve, which'd help a ton. Hell, using free field or diffuse field as a "gold standard" would also help a ton. Using ANYTHING as a reference point helps, always. Why is there so much opposition to setting a solid reference plane? Audio is an engineering effort, not magic. As I mentioned many times in different threads: the only way forward is via interactive sound customization towards the person. Anything else is nonsense, thus Harman would be a good point to drop the FR wars and go forward towards the better tomorrow full of customization instead of trying gear out in hope that something will fit what we want. Why buy and check if we can tweak and measure? The proper way to go is to make as many models for people with different head anatomy so anyone can find something that just fits and then modify the sound for what one wants, preferably in an automatic fashion. There is no need for anything else.

"Accuracy" is easy to define via the well understood suite of measurements for solid state devices, like DAC's and amplifiers - and is what pretty much everyone is looking for (with some exceptions). "Accuracy" is not easy to measure or quantify with transducers, whose measurements are represented by a 3-d sphere and have a time-domain, and the relationship between what is measured and what is heard is not completely understood. So by all means, continue "chasing accuracy" for transducers and see how far that goes.

The only way to chase accuracy with trancducers is to make ones that are possible to be tweaked to what is expected of them. Low-THD, bass-boosted reference point like Harman would serve helps mostly anyone, since people with different requirements would simply tweak the shelf to taste...

Sheesh, I can't understand the opposition towards standards. If not standards, the Internet would never work. It didn't before people invented the standarized TCP/IP stack...

All I really want to know is-- how do several good pairs of neutral loudspeakers with a flat direct response, good bass extension, and a fairly (though not necessarily perfectly) linear off-axis response measure at the eardrum when they're put into a typical semi-reflective domestic listening space.

See here. In case if Harman it was a far-field target, though, as not to account for variation in response due to moving in the speaker field.
AFAIR the actual measurements can be easily found in the net and I can't be arsed to find the specific paper right now.

The Harman Target Curve was based on a measurement at the primary listening seat in our listening room using Revel F208 and a JBL M2 calibrated to the Harman Speaker Target. We measured it with our ear simulated mounted in a head/torso. We did a spatial average +- 30 degrees by rotating the head.

 

[JJB70]

I would have no objection to the Harman curve or any other curve being used as a reference so that people could see the curve of a headphone relative to that reference. I am sure many would find that useful
My objection is the increasingly widely held idea that headphones tuned to the Harman curve are scientific, well engineered and correct, and others aren't. Other entities are quite at liberty to develop their own curves, which could be scientific or just what the designer likes. If that curve is well implemented with good technical performance then it has been well designed and engineered.

 

[preload]

Using Harman curves as a "gold standard" would help things, since anything else could be assessed in terms of deviance from a known common curve, which'd help a ton. Hell, using free field or diffuse field as a "gold standard" would also help a ton. Using ANYTHING as a reference point helps, always.

Help what? And who? Are you referring to a way to eliminate the Circle of Confusion?

Why is there so much opposition to setting a solid reference plane? Audio is an engineering effort, not magic. As I mentioned many times in different threads: the only way forward is via interactive sound customization towards the person. Anything else is nonsense, thus Harman would be a good point to drop the FR wars and go forward towards the better tomorrow full of customization instead of trying gear out in hope that something will fit what we want. Why buy and check if we can tweak and measure? The proper way to go is to make as many models for people with different head anatomy so anyone can find something that just fits and then modify the sound for what one wants, preferably in an automatic fashion. There is no need for anything else.

What you describe has already been done. AKG N90Q. It did auto EQ on your ears and had built in bass and treble tone controls. Overall, it sounded GOOD but NOT EXCELLENT. It also got discontinued without a follow up model, suggesting that it was a failed experiment from a profitability standpoint.

Sheesh, I can't understand the opposition towards standards. If not standards, the Internet would never work. It didn't before people invented the standarized TCP/IP stack...

Well, for starters, when talking about human subject preferences, there is always going to be variability in your measured data - with very few exceptions. That's like saying a standard male is 5'-10" and 70kg and designing everything around those parameters.

Also, a simple, smoothed, 2-d amplitude vs frequency curve (such as a "target curve") is not a complete characterization of what is "heard" by a set of headphones. In reality, what is being heard is a polar frequency response that is dependent on angle of incidence to the ear axis with micro time delays caused by reflections inside the ear cup, which represents a 4th dimension time). For this reason, you can't "EQ" an HD600 into an HD800S (and I'm pretty sure countless people have tried and were sorely disappointed).

 

[JJB70]

I had the N90Q briefly It was a strange design in some ways, sound quality was good but not as spectacular, certainly not for the price. I offloaded them because of quality issues, the experience of doing a warranty return and claim with Harman Europe (I bought them directly) put me off ever buying anything from them again.