What is the best headphone if HRTF and FR is accounted for?

[propaganda1]

Apologies for the weird title, I'm not good at wording things.

I recently watched a video from The Headphone Show where they got their HRTF's with specific headphones measured, linked below (I've timestamped to the relevant part).

So what I got from this video (please correct me if I am wrong) is that they now have information on how that specific headphone interacts with their head, and are able to accurately EQ (even in the treble) the response. Which got me thinking, does this mean theoretically if someone had access to this measuring system, they could make any headphone sound 'perfect' (Flat)? I'm trying to think of other important factors and have come up with the following:

- Distortion levels
- un-EQable resonances (I've read about this somewhere but am unsure if this is even a thing)
- Low positional variance (What causes this? Can it even be quantified?)
- The ability to play full range (e.g. the sennheiser open backs fall off in the subbass) (but this could fall under distortion levels)

Which comes to my question, what headphone would be the best if someone was to do this?

 

[Jiraya369]

to answer the title, it's the headphone that accounts for your head's FR/HRTF then lol

The positional variations are easily measured by putting the headphone forward, backwards, etc. Listener's IEM tests have him doing differing levels of insertion depth to better simulate how a random average person would wear it.

Unheard lab and Earphones Archive do really nice testing for positional variations. Here's an example

"This graph illustrates how headphone placement on the head affects perceived tonal balance: with the ear positioned at the front (blue), centre (green), and back (red) of the driver."


UL
EA

In general, the Sennheiser HD600 and the Edition XV are really good choices, the 600 falls off in subbass but not by much if you're an average person with average levels of distortion audibility. I'd say the DCA Noire X is the one that best targets the response for people as it is based on both the harman GRAS and the DF 5128 curves, it looks really nice on graphs. Just 1-3 eq bands and you're good to go.

You can eq treble FAIRLY accurately but don't expect too much. Without a point of reference it can be damn near impossible, i mean even with a reference you'd struggle to EQ.

Just look for comfort and a semi-decent FR, the rest can be eq'd pretty well with just a simple EQ to harman and additional tweaks to taste using a tone generator and turning down peaks.

 

[propaganda1]

to answer the title, it's the headphone that accounts for your head's FR/HRTF then lol

The positional variations are easily measured by putting the headphone forward, backwards, etc. Listener's IEM tests have him doing differing levels of insertion depth to better simulate how a random average person would wear it.

Unheard lab and Earphones Archive do really nice testing for positional variations. Here's an example


"This graph illustrates how headphone placement on the head affects perceived tonal balance: with the ear positioned at the front (blue), centre (green), and back (red) of the driver."


UL
EA

In general, the Sennheiser HD600 and the Edition XV are really good choices, the 600 falls off in subbass but not by much if you're an average person with average levels of distortion audibility. I'd say the DCA Noire X is the one that best targets the response for people as it is based on both the harman GRAS and the DF 5128 curves, it looks really nice on graphs. Just 1-3 eq bands and you're good to go.

You can eq treble FAIRLY accurately but don't expect too much. Without a point of reference it can be damn near impossible, i mean even with a reference you'd struggle to EQ.

Just look for comfort and a semi-decent FR, the rest can be eq'd pretty well with just a simple EQ to harman and additional tweaks to taste using a tone generator and turning down peaks.

Thanks for the reply, but I am a bit confused, if the headphone is measured on your head like in the video, wouldn't you be able to EQ treble (and the rest of the frequency) perfectly, seeing as you know exactly where the dips and peaks of that headphone are for your HRTF?

 

[kemmler3D]

Thanks for the reply, but I am a bit confused, if the headphone is measured on your head like in the video, wouldn't you be able to EQ treble (and the rest of the frequency) perfectly, seeing as you know exactly where the dips and peaks of that headphone are for your HRTF?

That's the theory, I guess, and for the sake of discussion let's take it at face value.

Since we are assuming FR is perfect for the user, I think the best headphone would have to then be the one with cleanest phase distortion / group delay and/or the one with the lowest overall distortion.

However, I do think (I read somewhere?) the angle / position of the drivers relative to the ear can influence spatial perception... if that is accounted for in the EQ, then maybe it's all about distortion.

I think the LCD-XC has really low distortion for an over ear (that's why I bought them) but I'm not sure if it's the lowest for your dollar.

 

[Jiraya369]

Thanks for the reply, but I am a bit confused, if the headphone is measured on your head like in the video, wouldn't you be able to EQ treble (and the rest of the frequency) perfectly, seeing as you know exactly where the dips and peaks of that headphone are for your HRTF?

if you have that data then you can eq ANY headphone

how will you yourself get the data for your own head?

 

[Jiraya369]

That's the theory, I guess, and for the sake of discussion let's take it at face value.

Since we are assuming FR is perfect for the user, I think the best headphone would have to then be the one with cleanest phase distortion / group delay and/or the one with the lowest overall distortion.

However, I do think (I read somewhere?) the angle / position of the drivers relative to the ear can influence spatial perception... if that is accounted for in the EQ, then maybe it's all about distortion.

I think the LCD-XC has really low distortion for an over ear (that's why I bought them) but I'm not sure if it's the lowest for your dollar.

noire x is cheaper with good distortion and an even better FR

i tried an lcd2 closed and i liked the spacious cups, i really like large planar headphones cus of that. The Openness leads to a larger sense of stage.

 

[propaganda1]

That's the theory, I guess, and for the sake of discussion let's take it at face value.

Since we are assuming FR is perfect for the user, I think the best headphone would have to then be the one with cleanest phase distortion / group delay and/or the one with the lowest overall distortion.

However, I do think (I read somewhere?) the angle / position of the drivers relative to the ear can influence spatial perception... if that is accounted for in the EQ, then maybe it's all about distortion.

I think the LCD-XC has really low distortion for an over ear (that's why I bought them) but I'm not sure if it's the lowest for your dollar.

Thank you for the reply

Are 'un-EQable resonances' a thing? A friend of mine told me about this a while but didn't really expand on it, if this does exist, can this be identified from FR graphs?

if you have that data then you can eq ANY headphone

how will you yourself get the data for your own head?

I may be able to visit the lab in the video soon, so I am planning to buy a single pair of headphones that will hopefully allow me to take a long break from the hobby

 

[solderdude]

Apologies for the weird title, I'm not good at wording things.

I recently watched a video from The Headphone Show where they got their HRTF's with specific headphones measured, linked below (I've timestamped to the relevant part).

So what I got from this video (please correct me if I am wrong) is that they now have information on how that specific headphone interacts with their head, and are able to accurately EQ (even in the treble) the response. Which got me thinking, does this mean theoretically if someone had access to this measuring system, they could make any headphone sound 'perfect' (Flat)? I'm trying to think of other important factors and have come up with the following:

- Distortion levels
- un-EQable resonances (I've read about this somewhere but am unsure if this is even a thing)
- Low positional variance (What causes this? Can it even be quantified?)
- The ability to play full range (e.g. the sennheiser open backs fall off in the subbass) (but this could fall under distortion levels)

Which comes to my question, what headphone would be the best if someone was to do this?

Planar magnetic headphones usually have the lowest distortion.
They usually also have a lower positional variance than dynamics which is not that important IF you measure them in the described way as you will put them on as you always would.
A question could be... is the used microphone (not EDRP) accurate enough ?
This method too is just in its infancy.

There is no need to visit a lab. You just need to own/use that specific (calibrated) in-ear mic + software.
This will probably cost as much a decent headphone but then you only have to buy a (at least decent) and comfortable headphone that suits your needs. (Closed/open/wireless)

if you have that data then you can eq ANY headphone

Only if you have that data for any headphone.
The thing is that the HRTF for different headphones will be different for each headphone. There isn't a single 'correct' correction that will work on any headphone.
This method measures the combination of ear+headphone.

Are 'un-EQable resonances' a thing? A friend of mine told me about this a while but didn't really expand on it, if this does exist, can this be identified from FR graphs?

Some headphones have long lasting resonances. Those usually are also visible in the frequency response as peaks OR dips.
Probably best to use noise for this type of measurements if you want to insure a dip is not overcompensated for FR.
You can't EQ away the time factor (ringing time) but can EQ away the FR part of it which is usually the biggest audible issue.
Resonances in planars usually are very narrow but can be very long lasting.
One should keep in mind that in music recordings all instruments have an attack and decay that usually lasts longer than those of transducers.

 

[Jiraya369]

Planar magnetic headphones usually have the lowest distortion.
They usually also have a lower positional variance than dynamics which is not that important IF you measure them in the described way as you will put them on as you always would.
A question could be... is the used microphone (not EDRP) accurate enough ?
This method too is just in its infancy.

There is no need to visit a lab. You just need to own/use that specific (calibrated) in-ear mic + software.
This will probably cost as much a decent headphone but then you only have to buy a (at least decent) and comfortable headphone that suits your needs. (Closed/open/wireless)

Only if you have that data for any headphone.
The thing is that the HRTF for different headphones will be different for each headphone. There isn't a single 'correct' correction that will work on any headphone.
This method measures the combination of ear+headphone.


Some headphones have long lasting resonances. Those usually are also visible in the frequency response as peaks OR dips.
Probably best to use noise for this type of measurements if you want to insure a dip is not overcompensated for FR.
You can't EQ away the time factor (ringing time) but can EQ away the FR part of it which is usually the biggest audible issue.
Resonances in planars usually are very narrow but can be very long lasting.
One should keep in mind that in music recordings all instruments have an attack and decay that usually lasts longer than those of transducers.

thank you for the truthnuke, i have been enlightened.

 

[kemmler3D]

Are 'un-EQable resonances' a thing?

Others have answered, but the short answer is yes. Resonances are a physical property of an object, and so you can't get rid of them by changing the input signal. You definitely want to avoid significant resonances when you're shopping for headphones.

One should keep in mind that in music recordings all instruments have an attack and decay that usually lasts longer than those of transducers.

True but this says nothing in particular for the audibility of a resonance, especially a narrow one. An extreme resonance would cause a broadband signal like a snare drum to sound like it also has a sharp tone or ringing sound. This doesn't need to last long at all to throw things off, it could be (say) just 30 milliseconds or so and still be easily audible.

 

[propaganda1]

Others have answered, but the short answer is yes. Resonances are a physical property of an object, and so you can't get rid of them by changing the input signal. You definitely want to avoid significant resonances when you're shopping for headphones.

Is there a way to accurately identify these resonances when looking at FR graphs? How can I know which peaks and dips can be EQ'd vs one that is a resonance of the headphone itself?

Solderdude mentioned noise, and I wanted to learn about other methods to accurately identify this.

I am considering an Audeze LCD-5, I have found a few used ones for much below the new price, but of course I am all ears to cheaper recommendations that would get me the same performance.

Thank you!

 

[solderdude]

Resonances can be spotted in the frequency response, phase response, group delay, oscilloscope shots, CSD (Waterfall).
They are caused by 'uncontrolled' vibrations and cannot (easily) be undone with EQ.
You can EQ the most audible aspect of a resonance namely the amplitude (frequency response).
A dip in the frequency response, however, can also be caused by a resonance and ring.

But mostly they also have a (narrow) peak.

I would not worry about those too much.

The worst offenders in ringing appear to be the planars. For some reason the least offenders appear to be the Audeze.

 

[kemmler3D]

Is there a way to accurately identify these resonances when looking at FR graphs? How can I know which peaks and dips can be EQ'd vs one that is a resonance of the headphone itself?

Solderdude mentioned noise, and I wanted to learn about other methods to accurately identify this.

I am considering an Audeze LCD-5, I have found a few used ones for much below the new price, but of course I am all ears to cheaper recommendations that would get me the same performance.

Thank you!

Yep as @solderdude mentioned, I find CSD to be the best way to spot resonances.

The ideal way such a graph looks is basically a flat wall with a tail to the left, those wavelengths inherently take longer to happen.

What you want to avoid are tall, narrow peaks/ridges that extend far out. Those are resonances. The height is how loud they are, and the length is how long they last. Higher and longer are both worse.

 

[solderdude]

There are people that absolutely love the hyper detailed /lively sound of the $5k headphone below.

Below the also well loved Edition XS

some find the headphone below boring:

 

[propaganda1]

There are people that absolutely love the hyper detailed /lively sound of the $5k headphone below.


Below the also well loved Edition XS


some find the headphone below boring:

To my noobie eyes, the last graph you posted looks like the best performing in terms of resonances, am I correct?

It seems that is the HD650 graph, which is a headphone I was considering but unfortunately it falls off in the sub-bass and is prone to distortion with heavier EQ, and there seems to be some resonances according to Amir's review of them. Out of all the CSD graphs you've seen, which headphone would you say had the least resonances?

Maybe I should make a new thread with a better title now that I know what I'm looking for lol

 

[solderdude]

Yes.... HD 650 if you are wondering 34mm dynamic driver.
Sennheiser was already able to make well measuring and sounding headphones in 1993 (HD 580 precision)

The plots are there to show to headphones that are really liked by audiophiles have a LOT of resonances yet are not found to sound 'bad' .... on the contrary even.

If you want good sound get something that is known to require little EQ, is comfortable and has low distortion in the bass at 104 and 114dB levels.
Above 1kHz you must only look at the 94dB plots (there are no 104dB peaks at these frequencies in loud music).
Also you need an amplifier that can drive your headphone to at least 120dB peaks even when one believes they are only 'listening at 70dB average'.

Very important .... comfort and for closed headphones ... seal.

 

[Budgie]

Ultimately, no substitute for listening yourself. My suggestion is in any choice you make, consider the return policy to be very important.

 

[Jiraya369]

Is there a way to accurately identify these resonances when looking at FR graphs? How can I know which peaks and dips can be EQ'd vs one that is a resonance of the headphone itself?

Solderdude mentioned noise, and I wanted to learn about other methods to accurately identify this.

I am considering an Audeze LCD-5, I have found a few used ones for much below the new price, but of course I am all ears to cheaper recommendations that would get me the same performance.

Thank you!

LCD-5 is nice but you're placing too much emphasis on resonances and stuff, they're narrow and dont affect your experience too much. If you can eq the large deviations you're good to go.

I've listened to an LCD-5, and for my HRTF it was damn near perfect actually so i loved it. But i eq'd my HE1000 Stealth and it was just even better. The Noire X has eq'able resonances and just has a near pitch perfect FR out of the box. First buy that then look at other things.

Ofc if you want to buy the LCD-5 cus of some magical things in it that's upto you. It certainly looks and feels nice but so do hifiman's headphones imo.