Headphones and the Harman target curve

[BrEpBrEpBrEpBrEp]

And they have since updated their curve.

A partial manuscript of the 2018 convention paper was peer reviewed. If you're really hell-bent on it, you can always stick with the 2015 curve. Or you can compare the methodologies yourself.

The truth is that there doesn't really seem to be a credible alternative target anyways.

 

[Yorkshire Mouth]

How would you define "accuracy" or "faithful transparency" in a measurable way? You mention "science," and science requires observability or measurability.




It's been done. And the theoretical FR curve measured at the ears that represents the FR of GSIAGR doesn't translate into the headphones that are accurate or preferred. Part of the reason is that a simple FR curve (and btw, you haven't even defined where you're taking the measurement from, i.e. in the canal, at the external auditory meatus, etc.) does not fully describe for what is heard. In reality, sound waves are striking the pinna at all sorts of different angles of attack, each with various millisecond delays and differences in amplitude with respect to one another. How, pray tell, could you possibly represent all of that with a single amplitude vs. frequency curve?



Again, you need to define how one should measure "accuracy." If you can't observe or measure it, it ain't science.

I think I’ve said this already.

Get trained listeners in the GSIAGR set up with headphones EQd as to what you think will make them sound the same. Then try it.

Have them play the same selection of material through the speakers and through the headphones.

DO NOT ASK WHICH THEY PREFER. Ask which headphone curve sounds closest to the room.

Again, I’m not saying that’s going to be perfect. But it should get you closer than asking for preference.

We already know that preference is ‘wrong’ if you’re aiming for accuracy. As soon as people started listening to GSIAGR, they turned up the bass, so we know we’re deviating from that target as soon as we start.

 

[BrEpBrEpBrEpBrEp]

As soon as people started listening to GSIAGR, they turned up the bass, so we know we’re deviating from that target as soon as we start.

We really, really all need to agree on the fact that GSIAGR includes boosted bass and recessed treble before discussing this. That downward slope from 20 Hz to 20 kHz? That's good speakers in a good room. Flat response from 20 Hz to 20 kHz? That's good speakers in an anechoic chamber.

 

[solderdude]

We already know that preference is ‘wrong’ if you’re aiming for accuracy. As soon as people started listening to GSIAGR, they turned up the bass, so we know we’re deviating from that target as soon as we start.

AFAIK the Harman research was done to educate the folks at Harman and to determine what 'tuning' customers prefer.
This was to increase sales no doubt.

This is something different than trying to come up with a new 'standard' which I would agree is the case. The problem is diffuse field and other current standards are even worse so in that light using Harman target (elevated bass and gently rolled off treble in practice) makes sense by lack of anything more suited not being available.

 

[BrEpBrEpBrEpBrEp]

AFAIK the Harman research was done to educate the folks at Harman and to determine what 'tuning' customers prefer.
This was to increase sales no doubt.

This shows through in the later papers especially, I believe. IIRC they remarked later on that untrained listeners preferred more bass and treble.

 

[Yorkshire Mouth]

AFAIK the Harman research was done to educate the folks at Harman and to determine what 'tuning' customers prefer.
This was to increase sales no doubt.

This is something different than trying to come up with a new 'standard' which I would agree is the case. The problem is diffuse field and other current standards are even worse so in that light using Harman target (elevated bass and gently rolled off treble in practice) makes sense by lack of anything more suited not being available.

I completely agree, and gave no criticism of Harman for that.

The annoying thing is that, at the same time as doing this and asking for a preference, they could also have checked which curve best matched GSIAGR.

My main issue is that this preference curve is treated as a true target. In Amir’s excellent equipment reviews, he treats Harman as a target in much the same way as he does flat frequency and distortion below audibility when reviewing DACs. No other curve is considered. Failure to come close to Harman ensures a ‘not recommended’ badge.

Subjective preference becomes objective fact.

There’s a degree of flexi-logic. If you note that Harman is a preference curve, supporters will half-defend it as unimpeachable, whilst simultaneously saying “hey, it’s just a preference”. And then go back to measuring headphones as pass/fail based on Harman.

Surely we can agree that it’s one or the other. And then stick to that.

 

[RHO]

My main issue is that this preference curve is treated as a true target. In Amir’s excellent equipment reviews, he treats Harman as a target in much the same way as he does flat frequency and distortion below audibility when reviewing DACs. No other curve is considered. Failure to come close to Harman ensures a ‘not recommended’ badge.

Subjective preference becomes objective fact.

While I do not follow you on other points, I have to agree on this.

 

[solderdude]

Amir sees the Harman target as valid because of A: the extensive research done and B: it matches his personal observations.
Especially the latter, given the always subjective evaluation, validates the research for Amir and most likely also for Oratory.

It's a choice that has been made and IMO makes more sense than diffuse field correction for speakers.
Until there is another standard more suitable for headphones on 1 particular test fixture which can be created for other fixtures as well the Harman one makes sense. Those that want measurements without the 'preference' part for the lows can just imagine a 'horizontal target' below 500Hz.

 

[Yorkshire Mouth]

Amir sees the Harman target as valid because of A: the extensive research done and B: it matches his personal observations.
Especially the latter, given the always subjective evaluation, validates the research for Amir and most likely also for Oratory.

It's a choice that has been made and IMO makes more sense than diffuse field correction for speakers.
Until there is another standard more suitable for headphones on 1 particular test fixture which can be created for other fixtures as well the Harman one makes sense. Those that want measurements without the 'preference' part for the lows can just imagine a 'horizontal target' below 500Hz.

I agree that I would use diffuse field, or anything else currently available.

I’d just be a bit more aware when pass/failing headphones based on what is currently preference. Too much distortion, etc, yes. I’d just like Amir to be a tad more liberal when it comes to frequency curve.

By the way, respect to everyone contributing, as this lively debate has not spilled over. Thank you.

 

[Jimbob54]

I agree that I would use diffuse field, or anything else currently available.

I’d just be a bit more aware when pass/failing headphones based on what is currently preference. Too much distortion, etc, yes. I’d just like Amir to be a tad more liberal when it comes to frequency curve.

By the way, respect to everyone contributing, as this lively debate has not spilled over. Thank you.

I think he is more concerned about evenness of response pre eq, and ability to correct EQ . Huge peaks/ troughs and jaggedy treble pre eq is likely to get an unfavourable panther even if after EQ it can be smoothed . I dont think absence of bass or deviation from target automatically garners a poor review pre EQ. I could be wrong.

To be honest, if someone has a pair of headphones they like stock, the best target for them is whatever their FR curve is. EG- mentally overlaying the HD600 curve onto the DUT chart might be more meaningful to some than the target curve- but of course then we dont have a common language for non HD600 owners.

 

[Yorkshire Mouth]

All fair and good points.

 

[RHO]

Amir sees the Harman target as valid because of A: the extensive research done and B: it matches his personal observations.
Especially the latter, given the always subjective evaluation, validates the research for Amir and most likely also for Oratory.

It's a choice that has been made and IMO makes more sense than diffuse field correction for speakers.
Until there is another standard more suitable for headphones on 1 particular test fixture which can be created for other fixtures as well the Harman one makes sense. Those that want measurements without the 'preference' part for the lows can just imagine a 'horizontal target' below 500Hz.

I still get the impression that his subjective observations pre- and post-EQ are heavily biased by what was measured. I'm not saying that he's lying about his subjective evaluation. I think his bias makes him judge headphones that deviate from the Harman target too much, to his eyes, as sounding bad to his ears.
He knows very well what a headphone that adheres to the target should sound like. And if it doesn't sound like that, by default it's bad, because he doesn't hear what he wants to hear (something resembling the Harman target curve).
I'm not saying that he's not allowed to have his own subjective target that he likes. He like the Harman target curve and I'm perfectly fine with that. I'm less OK with the way he represents it as a hard fact that any deviation from is objectively bad/wrong. The line is drawn too hard.
It's much easier to objectively judge a headphone on THD.

In contrast to amp and DAC measurements we can say that we know what we measure. We want 0 THD, 0 noise, 0 deviation from a flat frequency response, 0 IMD, ... Because objectively that is perfection. Subjectively some may not like that. But it's very clear that that is an subjective deviation from an objective optimum.
I'm not 100% convinced that the HTC is such an objective optimum. It's a subjective target that many like, because it represents something that comes close to good sound from speakers. It is not objectively the same as the sound from those speakers.

 

[solderdude]

One can always look at the raw measurements and see them as an additional data point and ignore subjective findings when they don't gel with your own ones in case you happen to disagree with Amir's evaluation of a headphone you own and know well.

One would expect Amir to have experience with listening tests and their pitfalls.

 

[Jimbob54]

This review https://www.audiosciencereview.com/...l-celestee-review-headphone.21917/post-727407 I think has elements of all the points raised here .

 

[RHO]

One can always look at the raw measurements and see them as an additional data point and ignore subjective findings when they don't gel with your own ones in case you happen to disagree with Amir's evaluation of a headphone you own and know well.

One would expect Amir to have experience with listening tests and their pitfalls.

For people that know how to read a FR curve and know what the target curve means that is surely an option.
My main gripe is that his subjective observations do not come across as such. And for many I would assume they are seen as objective truths about the sound of some headphones.

 

[solderdude]

They are objective truth to Amir.
One can say the same for any other reviewer out there.
Often reviewers don't agree about the same headphones either.

I think you touched the general issue about reviews in general.
Anyone that takes certain reviewer's opinions/measurements as gospel should realize that headphones specifically are very hard to review and personal preference of the reviewer cannot be avoided.
Rtings is the only one that really tried and there are many that don't agree with their objective stance and are missing, to them, vital or relevant info.

Headphone reviews, IMHO, should have both measurements and subjective evaluations.
Measurements have unavoidable errors in them, subjective evaluations are laced with personal preference.

 

[RHO]

They are objective truth to Amir.

I had to think about that a little but I think I agree with you.
We just have to keep that in mind when reading/viewing a headphone review here (and any other site).

 

[markanini]

This review https://www.audiosciencereview.com/...l-celestee-review-headphone.21917/post-727407 I think has elements of all the points raised here .

Yup, it track Harman well and Amir doesn't like it. There's also the Aeon closed RT which doesnt track Harman well which Amir recommends. The measurement couldn't have biased Amir.

 

[Earfonia]

I completely agree, and gave no criticism of Harman for that.

The annoying thing is that, at the same time as doing this and asking for a preference, they could also have checked which curve best matched GSIAGR.

My main issue is that this preference curve is treated as a true target. In Amir’s excellent equipment reviews, he treats Harman as a target in much the same way as he does flat frequency and distortion below audibility when reviewing DACs. No other curve is considered. Failure to come close to Harman ensures a ‘not recommended’ badge.

Subjective preference becomes objective fact.

There’s a degree of flexi-logic. If you note that Harman is a preference curve, supporters will half-defend it as unimpeachable, whilst simultaneously saying “hey, it’s just a preference”. And then go back to measuring headphones as pass/fail based on Harman.

Surely we can agree that it’s one or the other. And then stick to that.

I agree to this, Harman is a 'preference target' not a pass / fail target.
I've been researching of my own preference target curve for years base on IEMs that relatively sound balanced to my ears, yet found some IEMs that don't match my target yet sound really good and natural!

I guess at this time, since there is no absolute target curve for headphones and IEMs, we should respect individual preferences while being clear with our own preference.

 

[preload]

I completely agree, and gave no criticism of Harman for that.

The annoying thing is that, at the same time as doing this and asking for a preference, they could also have checked which curve best matched GSIAGR.

One slight problem with that, in case you haven't figured them out already:
How would you determine which Great Speakers to use to generate your curve? Also, did it occur to you that "great speakers" are generally regarded as "great" due to the listener preferences?

My main issue is that this preference curve is treated as a true target. In Amir’s excellent equipment reviews, he treats Harman as a target in much the same way as he does flat frequency and distortion below audibility when reviewing DACs. No other curve is considered. Failure to come close to Harman ensures a ‘not recommended’ badge.

I'm not sure Amir is basing his recommendations solely on headphone preference curves, is he?