Sennheiser HD 650/HD 6XX

[Sal1950]

Yes measuring the individuals ear response is the way forward.

So what’s a good approach to doing that at home?

You have to make a appointment with Dr Amir to have the data ports installed. A new program is available to interface with his AP analizer.

 

[pkane]

You have to make a appointment with Dr Amir to the the data ports installed. A new program is available to interface with his AP analizer.

USB or SPDIF? I'd prefer a more modern interface, so I don't have to keep upgrading.

 

[Grave]

A couple things.

1) Tyll is dumbass because he believes in a lot of insane nonsense like snake oil cables and badly designed DAC/amps being hifi, so I cannot take his opinions seriously.

2) Speakers are not practical for me because I am a failure still living in my parents house. I will be using open headphones for the foreseeable future, which is fine because they sound amazing to me.

 

[Guermantes]

A new program is available to interface with his AP analizer.

Are you sure that's targeting the correct orifice?

 

[Sal1950]

with his AP analizer.

Are you sure that's targeting the correct orifice?

I was wondering if anyone caught that. LOL

 

[svart-hvitt]

Many consider 800s neutral but they simply are not. They are significantly bright and demonstrably measure so. They almost seem to contradict the known ear response graphs. I honestly think they are horrid.

I don't think Designs are hunch oriented. They work around the known ear response. They vary however for all the reasons discussed.

No the choice of headphones would be easier if our anatomy was identical.

View attachment 14355

@BE718 , I still think you’re avoiding the elephant here. Let me explain:

Serious speaker producers agree on the main design goal: Flat and smooth in open space.

And they agree on this too: It’s futile to account for an individual’s body related transfer function.

This is, broadly speaking, how the serious speaker designer works. So in open space, good speakers measure the same.

Then comes the headphone designers, and what do we see:

=> Frequency response of different headphones, both across producers and from the same producer, vary wildly.

You argue that headphone producers account for the body related transfer function, but you don’t stop and ask the obvious question: Why don’t speaker producers account for an individual’s transfer function? Should there be different standards when making speakers and headphones? Double standards are not twice as good.

It could be that a speaker producer is scientific, but the evidence of this is absent. The evidence of absence is the wildly differing frequency responses from the same producer.

Regrettably, to me it looks as if science is absent in the headphones area. They can’t even get the frequency curve right.

 

[March Audio]

No you are missing the point.

Speaker designers may aim for a flat free field response but rarely acheive it due to the constraints already mentioned.

Speakers are not cans, they ain't strapped to your ears, the behaviour is different. The ear resonances need to be accounted for. They have got the FR curve roughly right, but designs vary and people vary. Hence FR varies and personal experience varies.

I'm not going to disappear down your philosophical rabbit hole with you, science is very much present in headphone and speaker design.

 

[svart-hvitt]

No you are missing the point.

Speaker designers may aim for a flat free field response but rarely acheive it due to the constraints already mentioned.

Speakers are not cans, they ain't strapped to your ears, the behaviour is different. The ear resonances need to be accounted for. They have got the FR curve right, but designs vary and people vary. Hence FR varies and personal experience varies.

I'm not going to disappear down your philosophical rabbit hole with you, science is very much present in headphone and speaker design.

There are headphones whose measurements vary little depending on how you wear them. Example: HD800, cfr. Hertsens’ measurements.

Why the ear needs to be accounted for in headphones - but not in speakers - I cannot understand. To me, it looks like double standards.

Headphones don’t come with a prescription for different ears. If headphone producers wanted to account for different ears, they could’ve made different versions of the same model or include EQ in the headphones. Instead, chaos reigns. Who profits on confusion?

If you can point me to research on headphones - and producers following this research documenting their research oriented design with measurements - it would be great.

 

[SIY]

Why the ear needs to be accounted for in headphones - but not in speakers - I cannot understand. To me, it looks like double standards.

Pinnae represent an insignificant perturbation on room acoustics (where they represent a vanishingly small proportion of room volume), but a significant one in headphone acoustics (where they represent a significant fraction of the enclosed space).

What bothers me, and maybe one of the reasons I dislike using headphones, is that what happens with a standardized pinna will not be the same as what happens with my ear. There's no agreement on the measurement method, much less the target acoustics. This of course has not stopped the creation of standards. The Harman papers, IMO, created more uncertainty.

 

[svart-hvitt]

Pinnae represent an insignificant perturbation on room acoustics (where they represent a vanishingly small proportion of room volume), but a significant one in headphone acoustics (where they represent a significant fraction of the enclosed space).

What bothers me, and maybe one of the reasons I dislike using headphones, is that what happens with a standardized pinna will not be the same as what happens with my ear. There's no agreement on the measurement method, much less the target acoustics. This of course has not stopped the creation of standards. The Harman papers, IMO, created more uncertainty.

I follow what you wrote. And your point on absence of «agreement on the measurement method, much less the target acoustics», resembles what I wrote.

However, I don’t understand what you mean by the «Harman papers...created more uncertainty».

If we look up @Floyd Toole ’s concluding comments on headphones, we find this:

«Conclusion: the best sounding headphones sound like good loudspeakers in a good listening room. Because that is where recordings usually originate, is anyone surprised? The good news is that it is possible to anticipate sound quality using the right measurements, interpreted in the right manner.»

So there seems to be a target, which is in principle the same («sound like good loudspeakers in a good room») as for speakers. And because there is a target («the good news») we can predict if a measured headphone sounds good or not. I guess everyone agrees so far, right?

So the source of disagreement may come when I use the absence of a target curve in headphones as evidence of absence of science in headphones design.

@BE718 said I was practicing sophistry when I asked about standards and evidence of science in headphones design (@BE718 wrote: «I'm not going to disappear down your philosophical rabbit hole with you, science is very much present in headphone and speaker design»). But is what I wrote very different from what @Floyd Toole has commented? Let me quote Toole:

«Headphone measurements have been and still are a debated issue. (...) Then there is the second issue: what is the target curve for the measurement? (...) There was reason to believe that this kind of measurement, used with this kind of target, could be useful in designing headphones that have a chance of sounding good. But such was not to be the case. A small number of brands took the hint, and some visited me at the NRCC, but most companies apparently had other “marketing” priorities. Headphone sound has continued to be characterized by large variations».

In other words: As speakers have converged to a target (flat smooth curve), there has been no such convergence in headphones. How can audio be scientific if there’s absence of convergence? Is the headphones area still stuck in «marketing priorities»?

 

[SIY]

However, I don’t understand what you mean by the «Harman papers...created more uncertainty».

Yet another "target curve," which also varies as a function of measurement fixtures.

 

[pkane]

Pinnae represent an insignificant perturbation on room acoustics (where they represent a vanishingly small proportion of room volume), but a significant one in headphone acoustics (where they represent a significant fraction of the enclosed space).

What bothers me, and maybe one of the reasons I dislike using headphones, is that what happens with a standardized pinna will not be the same as what happens with my ear. There's no agreement on the measurement method, much less the target acoustics. This of course has not stopped the creation of standards. The Harman papers, IMO, created more uncertainty.

I fully agree with this. I've now tested flat, Harman-recommended, and simple (speaker) room-correction curves with two different headphones. I do prefer the Harman-recommended FR with a bit of an adjustment (3.5KHz bump rather than 4KHz) and a much lower increase of only about 4-5dB.

But what is strange to me is that there's no recommended way to determine the FR curve for my ears (except for @Sal1950 "analizer", of course!) If this is so important for headphone usage, why is there no software or hardware to create a DSP profile for my own ears? Or is there?

 

[Grave]

svart-hvitt, please stop posting nonsense.

Headphones are not speakers. There is a target frequency response for headphones developed by Harman. Google it. I think that this frequency response is roughly correct. The Harman target frequency response curve is somewhat similar to the Sennheiser HD 650's frequency response.

The Harman target frequency response: There is a dip at 200 Hz, followed by a peak at 3 kHz due to the shape of your ear, and everything beyond this is rolled off due to the lack of high frequency attenuation in headphones which would come from the channels interacting in speakers.

My equalized Sennheiser HD 650's: There is flat bass, followed by a peak at 3 kHz due to the shape of your ear, and everything beyond this is rolled off due to the lack of high frequency attenuation in headphones which would come from the channels interacting in speakers.

 

[Grave]

But what is strange to me is that there's no recommended way to determine the FR curve for my ears (except for @Sal1950 "analizer", of course!) If this is so important for headphone usage, why is there no software or hardware to create a DSP profile for my own ears? Or is there?

I believe this is something in development right now. You may be able to find something like this. I do not see the point though, you might as well just try different headphones and find the headphones which sound about right to you, and then equalize them.

 

[pkane]

I believe this is something in development right now. You may be able to find something like this. I do not see the point though, you might as well just try different headphones and find the headphones which sound right to you.

I'd want to EQ my headphones to a FR curve that works best with my ears. Trying out different headphones until I find one that happens to fit me exactly is a non-starter. If you have a pointer to any ear measurement software/hardware, please share.

 

[andreasmaaan]

Headphones are not speakers. There is a target frequency response for headphones developed by Harman. Google it. I think that this frequency response is roughly correct. The Harman target frequency response curve is somewhat similar to the Sennheiser HD 650's frequency response.

My equalized Sennheiser HD 650's: There is flat bass, followed by a peak at 3 kHz due to the shape of your ear, and everything beyond this is rolled off due to the lack of high frequency attenuation in headphones which would come from the channels interacting in speakers.

The Harman target is measured with reference to a dummy head and pinna, i.e. the headphones they used in the study were measured on their dummy, and the target response was calibrated based on measurements taken on this dummy. They've developed their pinna with a lot of care, but nevertheless their target response will not be correct for you if your head and pinnae are significantly dissimilar to Harman's.

Also, the Harman target curve is not flat in the bass. There's actually a 4dB shelf, which is lacking in the HD650s. These would be considered lean headphones with reference to the Harman curve.

EDIT: OE2017 in the graph below is the current Harman curve for over-ear headphones I believe:

 

[andreasmaaan]

There's a neat summary of how Harman developed their curve here.

 

[maverickronin]

There's a neat summary of how Harman developed their curve here.

Are those full papers available for free anywhere?

Asserting it's possible to test most headphones blind given that a person actually has to wear them seems pretty silly. There are a model families within the same manufacturer with identical or very similar fits that this could work for but the models on page 12 of that PDF have vastly different fits.

I go to meets and audio shows to try all the different models I can and could probably identify the make or model or plenty of headphones just by fit.

 

[andreasmaaan]

Are those full papers available for free anywhere?

Asserting it's possible to test most headphones blind given that a person actually has to wear them seems pretty silly. There are a model families within the same manufacturer with identical or very similar fits that this could work for but the models on page 12 of that PDF have vastly different fits.

I go to meets and audio shows to try all the different models I can and could probably identify the make or model or plenty of headphones just by fit.

I think just a few of the papers are available online, but most aren't. Try googling them by name and writing pdf in the search, I think I found a few that way.

What does being able to feel the headphones on your head have to do with the blind testing? You are still "blind" as to the brand, model and type of transducer(s), i.e. this method of testing still controls every other important variable (and since headphones can't be listened to unless on the head, this variable cannot be controlled).

 

[maverickronin]

What does being able to feel the headphones on your head have to do with the blind testing? You are still "blind" as to the brand, model and type of transducer(s), i.e. this method of testing still controls every other important variable (and since headphones can't be listened to unless on the head, this variable cannot be controlled).

I'm saying there are headphones which I could identify the make or model of just by wearing them.

Even if someone doesn't have that experience with which to identify specific models, having to wear them will still bias the result. Differences in comfort and fit will bleed over into the evaluation even if you tell people to only rate preferences in sound. I think the fact that this variable is not controlled for (even if it's impossible to do so...) should at minimum add some gigantic error bars and probably invalidate the whole thing.

I think for this to be valid it would need to be done with sets headphones which are essentially identical in fit and some of these do exist. The Sennheiser HD600/HD650, Beyerdynamic DT880/DT990, AKG K601/701/712 and much of the Stax and Audeze lines would be good sets to base tests around.