OLLO S5X Headphone Review

[Robbo99999]

I have both the HD560S and S5X (and many other headphones). The S5X simply is the better headphone. I can't make any more or less of it.

For studios there is the possibility to fine-tune the headphone sound which makes use of individual measurements of that particular headphone (S5X v1.2) in the DAW where simulations exist for certain studios.

It's true we have to draw our own subjective opinions of headphones. I place more faith in measurements vs a persons reported experience though. Having said that K702 is one of my most favourite headphones, and it doesn't measure too well - albeit I used measurements to optimise it for channel balance along with one of my K702 being measured by Oratory gave me a unique opportunity to be very certain about what the measured frequency response of my actual unit was (removing unit to unit variation variable) - so therefore with EQ to Harman Curve it is about as valid as can be in terms of knowing what measured frequency response you have for your particular unit of headphone. It's probably one of my most favourite headphones likely because all those variables of variation have been ironed out, combined with being large comfortable earcups which seem to create a large soundstage, and measured distortion is not great but ok at the volumes I listen at. K702 has also been shown to have low variation when worn from person to person in terms of the headphone coupling being quite reliable from one person's anatomy to the next. I think what I'm getting at is that frequency response that is actually received at the listeners eardrum is the most important thing, and there are so many factors that can prevent that from being reliable/predictable that it can be a bit of a crap shoot with headphones (when comparing them) - to some extent I think a person has to luck out with the frequency response that they actually receive at their ears when they're talking about the best headphone they possess. I mean I've got 4 units of K702 and I've measured all of them and channel balanced all of them with EQ and EQ'd them all to Harman Curve and I still think there's one of the units that sounds the best to me! So I think there's an element of luck involved, both in choosing the model of headphone that works best with your anatomy in terms of comfort and also the unique coupling that it has to your anatomy which will influence the frequency response that ends up at your eardrum, but also luck within the model of headphone you've chosen in terms of unit to unit variation.....which is also why it's good to choose a headphone that is likely or proven to some degree to have low unit to unit variation. This is all with focus on absolute fine tuning of sound for best possible experience, but it's also true that every headphone improves with EQ to Harman in my experience (as that's my preference for the target), just that to really fine tune your sound for best experience does seem to require a bit of "luck"! This has repercussions when people compare their own headphones and cite their favourites (especially if they have all been EQ'd to supposedly the same curve).

EDIT: With all of that previous paragraph in mind as a disclaimer/tamer/perspectifier(lol), the New Version HD560s frequency response would be better for me than what's listed here with Ollo.

 

[solderdude]

With all of that previous paragraph in mind as a disclaimer/tamer/perspectifier(lol), the New Version HD560s frequency response would be better for me than what's listed here with Ollo.

For you yes, I prefer the slightly modified S5X as the sound quality is higher. A bit like this part you wrote:

still think there's one of the units that sounds the best to me!

Could be real, could not be. In any case it is just an opinion, like mine is, without any comparative measurements that show it is different enough in certain measured performance which would be audible.
Could be as simple as different foam being used or product tolerance or silent change in the driver for instance.

The proof of the pudding is in the eating. When you want to be as sure as me it would require owning both the HD560S and S5X at the same time. Otherwise it is just an assumption based on measurements and/or opinions of others.

 

[Robbo99999]

For you yes, I prefer the slightly modified S5X as the sound quality is higher. A bit like this part you wrote:

Could be real, could not be. In any case it is just an opinion, like mine is, without any comparative measurements that show it is different enough in certain measured performance which would be audible.
Could be as simple as different foam being used or product tolerance or silent change in the driver for instance.

The proof of the pudding is in the eating. When you want to be as sure as me it would require owning both the HD560S and S5X at the same time. Otherwise it is just an assumption based on measurements and/or opinions of others.

Yes, but I'd like to make the point I'm more swayed to the measurement side generally, (certainly to get me in the right ballpark initially).

 

[Smaestro]

I bought the S5X with the optional "Unit-specific calibration" (USC, €49), here are my thoughts on that.

What is USC
What that means is that they measured this particular example when it was still in their hands, then shipped it out to me, and gave me a download link for a calibration file (in .xps format), and I got a second download link for a VST3 plugin in which to load that calibration file.

The big difference with, for example sending the headphones to any of the measures, and then correcting the EQ yourself, is that this is calibrated towards their S5X-target curve, and not to flat/harman/whatever else. Their S5X target curve is shown in smoothed pictures on their website, but AFAIK not in a Auto-EQ style file format.

Sound
I gave it a test today, by comparing it on/off on both a client song and pink noise.

I really didn't expect much, but two things stood out:
- Tonality was better. The main thing is that it adds a bit of bass in and cuts the mids, which is nice. As discussed above I like flat to work with, but it was a tad too flat.
- Channel matching was significantly better. It wouldn't be something I'd miss if I didn't have it, but now that I have, this plugin will not leave my monitoring chain. The best way to describe it is is that it adds the focus that listening in mono does. Without actually being mono of course.

Measurements of the USC file
(So this only applies to the file supplied to me for my headphones, so it does not apply to anyone else. Just sharing out of interest)
Lol measurements. Don't shoot me, I don't do measurements, but I was curious.

Frequency Response
EQ adjust (on Audiocheck mono pink noise on 3dB/oct slope analyser), white is normal pink noise without USC, red is with USC.

My specific USC adds a gradual 1dB bass boost, up to about 150Hz, from there a 1.2dB mid cut, until a crossover at 4kHz where highs get a 1dB boost.
Nothing weird.

I already thought about adding a 1-2dB bass boost, and this is basically what they thought too.

Channel matching
So with that massive change in perceived "focus", I expect big changes in the channel matching. Input was mono pink noise, output after USC plugin: Blue is left, Red is right.

Look at that enormous...! wait... whaaat? The difference between left and right is narely anything!
At some points Left is 0.1dB higher than the right, on other areas they are equal. Not what I expected at all.

So yeah.. Perhaps the slightly scooped frequency response is the cause of the better focus. Since a lot of bass content is mono, this curve literally adds more mono signal.
I could perhaps experiment with perceived width and bass level some day, especially since those are my trouble areas, it's interesting to see a possible relation.
But enough playing for today, need to work.

Other stuff supplied with the USC
The measurement chain is nicely laid out in a supplied document, and as I'm sure some of you are interested in that, so here it is:

• Ear sim: G.R.A.S. 45CC - IEC 60318-1, Couplers: RA0039, DAQ: Dewesoft SIRIUS MINI
• Calibration is done with pink noise crest 9.6 at 85 dBSPL

And then it goes on to describe another chain, it's not entirely clear to me with the verbose text, but it seems to refer to the rig with which the target was developed

• Ear sim: GR.A.S. KEMAR - IEC60318-4, Couplers RA0402, BK 5000/1 pinna, which was then transferred to IEC60138-1 ear sim.

I'm sure this stuff is on the website too and better explained (I don't know this stuff).

 

[Robbo99999]

I bought the S5X with the optional "Unit-specific calibration" (USC, €49), here are my thoughts on that.

What is USC
What that means is that they measured this particular example when it was still in their hands, then shipped it out to me, and gave me a download link for a calibration file (in .xps format), and I got a second download link for a VST3 plugin in which to load that calibration file.

The big difference with, for example sending the headphones to any of the measures, and then correcting the EQ yourself, is that this is calibrated towards their S5X-target curve, and not to flat/harman/whatever else. Their S5X target curve is shown in smoothed pictures on their website, but AFAIK not in a Auto-EQ style file format.

Sound
I gave it a test today, by comparing it on/off on both a client song and pink noise.

I really didn't expect much, but two things stood out:
- Tonality was better. The main thing is that it adds a bit of bass in and cuts the mids, which is nice. As discussed above I like flat to work with, but it was a tad too flat.
- Channel matching was significantly better. It wouldn't be something I'd miss if I didn't have it, but now that I have, this plugin will not leave my monitoring chain. The best way to describe it is is that it adds the focus that listening in mono does. Without actually being mono of course.

Measurements of the USC file
(So this only applies to the file supplied to me for my headphones, so it does not apply to anyone else. Just sharing out of interest)
Lol measurements. Don't shoot me, I don't do measurements, but I was curious.

Frequency Response
EQ adjust (on Audiocheck mono pink noise on 3dB/oct slope analyser), white is normal pink noise without USC, red is with USC.
View attachment 366642
My specific USC adds a gradual 1dB bass boost, up to about 150Hz, from there a 1.2dB mid cut, until a crossover at 4kHz where highs get a 1dB boost.
Nothing weird.

I already thought about adding a 1-2dB bass boost, and this is basically what they thought too.

Channel matching
So with that massive change in perceived "focus", I expect big changes in the channel matching. Input was mono pink noise, output after USC plugin: Blue is left, Red is right.

View attachment 366641
Look at that enormous...! wait... whaaat? The difference between left and right is narely anything!
At some points Left is 0.1dB higher than the right, on other areas they are equal. Not what I expected at all.

So yeah.. Perhaps the slightly scooped frequency response is the cause of the better focus. Since a lot of bass content is mono, this curve literally adds more mono signal.
I could perhaps experiment with perceived width and bass level some day, especially since those are my trouble areas, it's interesting to see a possible relation.
But enough playing for today, need to work.

Other stuff supplied with the USC
The measurement chain is nicely laid out in a supplied document, and as I'm sure some of you are interested in that, so here it is:

• Ear sim: G.R.A.S. 45CC - IEC 60318-1, Couplers: RA0039, DAQ: Dewesoft SIRIUS MINI
• Calibration is done with pink noise crest 9.6 at 85 dBSPL

And then it goes on to describe another chain, it's not entirely clear to me with the verbose text, but it seems to refer to the rig with which the target was developed

• Ear sim: GR.A.S. KEMAR - IEC60318-4, Couplers RA0402, BK 5000/1 pinna, which was then transferred to IEC60138-1 ear sim.

I'm sure this stuff is on the website too and better explained (I don't know this stuff).

How are you measuring your headphones, what are you using? I'm assuming the calibration files they provide also match the two channels together, which is why you're showing perfect channel matching through all the frequency response which is good (I essentially do the same using a miniDSP EARS measurement rig for all my headphones, the channel matching). Regarding their target curve I can't really comment much on that, except that it's not based on published research as far as I know, and validity for that target curve is therefore low. Given the changes the EQ is doing, it's making it a little bit more Harman like in terms of bass, but it's reducing pinna gain even more at around 2kHz which is where Ollo already measures very low vs Harman, which would almost end up with zero pinna gain after that EQ, which is sort of laughable (EDIT: if being unkind).

 

[Smaestro]

How are you measuring your headphones, what are you using?

I'm not measuring the headphones at all. I'm only measuring supplied VST plugin with the custom EQ profile, because the curve isn't visible in their plugin.

Regarding their target curve I can't really comment much on that, except that it's not based on published research as far as I know, and validity for that target curve is therefore low.

As far as 'validity' goes, it is has both up- and downsides to Harman.
On the one hand, Ollo's research is unpublished, so it could be faked indeed.
On the other hand it is (claimed to be) researched specifically on the ability for work in for spatial audio by audio professionals, instead of for enjoyment by rando's.
How you weigh these pro's and cons is up to you.

Given the changes the EQ is doing, it's making it a little bit more Harman like in terms of bass, but it's reducing pinna gain even more at around 2kHz which is where Ollo already measures very low vs Harman, which would almost end up with zero pinna gain after that EQ, which is sort of laughable.

You incorrectly compare the compensation curve for my specific headphone, to the FR of the one that Amir measured and conclude that it will lead to even less pinna gain. But perhaps my headphone had a lot more of that, and this curve compensates for that. You can't know.

Second, I don't know what you mean by laughable. Have you worked in spatial audio (the use case for these headphones)? Spatial audio itself is a big proprietary mess with Apple Sony and Dolby are doing their own thing, and thus the circle of confusion will completely repeat. But since spatial audio uses about 48 transfer curves instead of 1, it is easily defendable that a flatter headphone will be more useful to plot HRTF's on.

On a personal note. I don't understand the concept of "Pinna gain". So the idea is that when our ears hear the perfect anechoically-flat-speakers-in-a-normal-room, that the pinna adds a bit of 1-4khz, right? But these S5X are complete over ear headphones. The pinna is completely free from contact from the pads. So the pinna is not limited in their ability to "gain". Why should the headphones have a boost?

 

[solderdude]

Only the concha gain, pinna flange gain and ear canal gain remain with over ear headphones.

Concha gain peaks around 5kHz but is very directional and since sound comes from the side it is less 'effective' so there is less gain there.
That gain is present with speakers (as is torso, neck and spherical head effects)
Note the drawing below is for sounds coming from an angle of 45 degrees (headphones is 0 degrees and differs) but shows the parts of the frequencies it 'boosts' the signal

Pinna flange gain is highest around 3.5kHz and also directional and not really present with headphones as sounds do not come from the front.

The ear-canal gain is always present and does not seem to be directional (the pinna does that) It is a tube that likes to resonate around 2.5kHz.
It is why the hearing is most sensitive in that part of the frequency range. We don't perceive sounds as boosted in that area though as the brain takes that in consideration (calibrates itself to real world sounds).
The ear-canal gain is rather big and somewhere between 2-3kHz and differs from person to person as length, diameter and shape differ.
Its there where it turns out not everyone hears the same as a standard test fixture and people want deviation from standards in practice.
It is also why many headphones (quite obvious in Sennheiser headphones) seemingly have a dip around that part of the frequency range opposite test fixtures.

So depending on the ones ear gain the brain 'calibrates' itself to surrounding sounds coming from in front of them. This differs from sounds coming from the side.
Sounds coming from the side interact differently with the ear than sounds coming from the front (2-ch audio) through the air and hear 2 sound sources at the same time.
Harman research is based on the perception of speakers in a 'listening room' coming from the front and attempts to create a similar tonal balance for headphones so it mimics speakers in a room.

With multichannel sound there is also sound coming from other directions. The S5X is designed to be close to that soundfield but only when mixed down to 2 ch.
This differs from Harman goals and thus the target differs.

Ollo has done extensive measurements in various studios and worked with them to create the headphones. They are not intended for music enjoyment but can be used for that too, just not Harman target. Some EQ will get them there of course.

 

[Robbo99999]

I'm not measuring the headphones at all. I'm only measuring supplied VST plugin with the custom EQ profile, because the curve isn't visible in their plugin.


As far as 'validity' goes, it is has both up- and downsides to Harman.
On the one hand, Ollo's research is unpublished, so it could be faked indeed.
On the other hand it is (claimed to be) researched specifically on the ability for work in for spatial audio by audio professionals, instead of for enjoyment by rando's.
How you weigh these pro's and cons is up to you.




You incorrectly compare the compensation curve for my specific headphone, to the FR of the one that Amir measured and conclude that it will lead to even less pinna gain. But perhaps my headphone had a lot more of that, and this curve compensates for that. You can't know.

Second, I don't know what you mean by laughable. Have you worked in spatial audio (the use case for these headphones)? Spatial audio itself is a big proprietary mess with Apple Sony and Dolby are doing their own thing, and thus the circle of confusion will completely repeat. But since spatial audio uses about 48 transfer curves instead of 1, it is easily defendable that a flatter headphone will be more useful to plot HRTF's on.

On a personal note. I don't understand the concept of "Pinna gain". So the idea is that when our ears hear the perfect anechoically-flat-speakers-in-a-normal-room, that the pinna adds a bit of 1-4khz, right? But these S5X are complete over ear headphones. The pinna is completely free from contact from the pads. So the pinna is not limited in their ability to "gain". Why should the headphones have a boost?

Ah right, ok, (you didn't measure them), so you have no idea if the headphone is channel matched well or not, at least not independently verified, so I don't think you can really claim that the headphone is channel matched through the whole frequency range.

You'd hope that unit to unit variation was not large so I applied the EQ for your individual headphone to Amir's measurement, but yes we don't know. But this point would also potentially negate (make less likely) your idea that the channels are matched perfectly.

Re "laughable", I put an EDIT just now into my previous post, I was being unkind, I knew that, but I found it pretty surprising (& amusing) they wanted to reduce that what is often called pinna gain area even further given that was one of the most significant deviations of this headphone away from ye olde Harman Target. I don't think your points re spatial audio and perhaps misuse or confusing use of the word "flatter" are particularly helpful or true to task. The rise in response with peak at around 3kHz for the Harman Target is about simulating speakers in a room tonally as solderdude posted above (although those dB levels seem a bit inflated in his graph in relation to Harman Target for instance). In the Harman Research (2018 over ear) it's effectively anechoic flat speakers in a good listening room measured at the "ear drum" of a dummy head at listening position with a bit more bass and a tad more treble added on, so when you measure headphones on a dummy head (GRAS) then you want it to follow that target curve to simulate those speakers, because of course when you measure the headphones you're measuring at the eardrum of the dummy head too - it's about matching it at the eardrum measurement (which is why it's not a visually flat line) to simulate good speakers in a good room.

EDIT: @solderdude , I like the fact that they used measurements and had a theory for what they were trying to achieve, re the studio development work you say they did.......I'm unsure how valid it is though. In my headphone EQ experience it's not a target that I'd like, but I don't want to dwell on it too much, because I've been pretty negative about it over my last few posts.

 

[Smaestro]

Very interesting, thanks.

Only the concha gain, pinna flange gain and ear canal gain remain with over ear headphones.

Only? Those are the 3 most important ones in the graph.
Becaues the "Torso and Neck" does nothing above 2kHz, and infact dips before that, so removing that adds to the gain in 1-2kHz.
The head is a smooth shelf, so that is more of a "tilt" than a EQ profile. Seems to me really that the ear is more than capable to create the same bell peak of gain as with speakers.

Pinna flange gain is highest around 3.5kHz and also directional and not really present with headphones as sounds do not come from the front.

Or not.. haha. But still the Concha and Ear canal remain and they are the dominant 'curves' in that image.

The Ollo do have offset drivers.
Let alone the driver angle of 6 degrees, putting the driver center more forward, would change the angle of the incoming audio more towards the front. Assuming that equals a 6 degrees, that's 13% of the way to 45 degrees, so that might add a little of the pinna flange gain back.

The ear-canal gain is rather big and somewhere between 2-3kHz and differs from person to person as length, diameter and shape differ.
Its there where it turns out not everyone hears the same as a standard test fixture and people want deviation from standards in practice.
It is also why many headphones (quite obvious in Sennheiser headphones) seemingly have a dip around that part of the frequency range opposite test fixtures.

The infamous BBC Dip

Gundry Dip or BBC Dip?

Gundry Dip or BBC Dip?

hydrogenaud.io

Interestingly, multiple friends have commented on that the S5X sound the most like sounding like speakers.

 

[Smaestro]

Ah right, ok, (you didn't measure them), so you have no idea if the headphone is channel matched well or not, at least not independently verified, so I don't think you can really claim that the headphone is channel matched through the whole frequency range.

You'd hope that unit to unit variation was not large so I applied the EQ for your individual headphone to Amir's measurement, but yes we don't know. But this point would also potentially negate (make less likely) your idea that the channels are matched perfectly.

Hmm, I don't think I claimed that they were channel matched. Either way, I certainly didn't intend to do so. What I said (or think I said) was:

They subjectively sounded much more focused to me after applying the EQ curve, which I attributed to corrective channel-matching.

However, after analysing what the EQ actually does, I showed that it does barely anything (0 or 0.1 dB between left and right, close to nothing). With that I tried to say that my initial thought must have been wrong, and that no channel matching happened, or at least not "as immensely" as I experienced it. I then suggested another explanation for this effect, namely that a higher level of (mono) bass leads to a more focused perception.

But all of this is subjective, so if I seemed to implied that my S5X actually is channel matched, or that I could accurately determine that by just listening, then I'm sorry for that, that was unintended.

In the Harman Research (2018 over ear) it's effectively anechoic flat speakers in a good listening room measured at the "ear drum" of a dummy head at listening position with a bit more bass and a tad more treble added on, so when you measure headphones on a dummy head (GRAS) then you want it to follow that target curve to simulate those speakers, because of course when you measure the headphones you're measuring at the eardrum of the dummy head too - it's about matching it at the eardrum measurement (which is why it's not a visually flat line) to simulate good speakers in a good room.

I am a bit confused about what "the Harman research" actually says about the over-ear curve. In some posts it is described in the way you said (flat speakers in good room, measured in model ears). In some posts it is described as users twisting an EQ until the perception of the headphones was equal to that of the flat speakers in a good room.

I did search ASR and other sites for 'the truth', but there's just sooo soooo much written about all this that I can't find the forest for the trees.

With multichannel sound there is also sound coming from other directions. The S5X is designed to be close to that soundfield but only when mixed down to 2 ch.
This differs from Harman goals and thus the target differs.

Ollo has done extensive measurements in various studios and worked with them to create the headphones. They are not intended for music enjoyment but can be used for that too, just not Harman target. Some EQ will get them there of course.

How is a surround field be mixed down to 2 ch? I mean, the mix down in production is clear: you are wearing your Apple Vision and you look in a certain direction, the immersive mixer gives every sound source a HRTF suitable to the location of the source.

But the S5X is a static headphone, no active mixing down. What does it mean in this context to mix a surround field down to 2ch? Has a dummy head been placed in a room where it is hit by audio equally from every direction, and then measure the FR at the dummy eardrum?

 

[Smaestro]

I found the forest, ahem, the overview of the research I was looking for:

The Perception and Measurement of Headphone Sound Quality- What Do Listeners Prefer? - Sean E. Olive

acousticstoday.org

 

[solderdude]

How is a surround field be mixed down to 2 ch?

By mixing the channels in a certain ratio (and FR corrected).

Has a dummy head been placed in a room

Yes, measurements were made in various studios using a dummy head to measure the sound field.

 

[Smaestro]

What I don't understand then:
What does it mean that the S5X has a frequency response optimised for mixing spatial audio? If the S5X FR is the result of some sort of average of all sounds in the Dolby studio's, measured in the ear, doesn't it mean that, then theoretically the headphone always applys a sort of average HRTF (EQ only) of all directions? That would sounds as much as possible as sound coming from anywhere / everywhere / indeterminate. Not what it sounds like to me though.

Where am I going wrong here?

What I do understand:
When actually using fully spatial audio (Apple Vision, or anything motion tracking basically), the device will mix the virtual audio sources, give them an appropriate loudness, FR and time delay, to simulate their HRTF, and then the user gets the best illusion of spatial / full surround audio. Surely with a angled speaker array this can be more accurate, but if each ear gets only one speaker, then everything gets mixed to 2 channel. When a song is mixed to 'spatial audio', and played without head movement tracking, IMO that's just a bunch of EQ, transient and Haas delay 'tricks' to make the track sound fuller than normal. This is something that mixing engineers do anyways, from long before it was called spatial audio.

The 'proof' of this is in trying out the headphone Atmos test on the Dolby website. A strange experience IMO. You can switch between Atmos and traditional Stereo, and of course they made the Stereo signal so much lower in volume that the Atmos version sounds so much better... https://www.dolby.com/atmos-visualizer-music/
But if you level match, you find it's just a more spatiously mixed version of the song. There are many indie singer-songwriters with very densely packed yet spacious songs, but they don't called it spatial audio. Spatial audio in static 2 channels is just..... stereo audio with a fancy name.

But to not be too critical, it is good that the mixing workflow gets new tools. Instead of every mix engineer figuring out for him/herself how to achieve more dimensionality, mixing tools improve to place anything wherever you like in a more straight forward way.

 

[markanini]

I don't know enough of the research to claim the following, so this may be completely wrong. But here's why I think the Harman curve masks more than the equivalent live room situation:

First the research:
1. It was determined that most listeners prefer a flat speaker response in a good room with smooth selectivity.
2. It was deterimined that most listeners find the Harman target curve most neutral or equivalent to this flat room response.

But at no point is tested wether individual parts of the music were equally intelligible.

Given that:
- a possible explanation for the curves existance exists (compensation for the lack of bodily bass sensation),
- and that listeners were not instructed to pay attention to masking,
- and it was done with untrained listeners who could not do that even if instructed

I conclude that the headphone curves can not be used to assume anything about equal masking, and therefore the general theory about masking still stands.

As an extra, Harman target also has elevated highs. Research shows that loud tones over 1kHz mask more towards lower freqencies than higher. So the mids get masked from both sides.

I end this again with: I am still beginning on reading up on all the research and it's quite possible that any or all of the above is wrong.

I recommend reading the blog of Sean Olive chronologically, earliest posts first. I'ts rather easy to digest and explains the motivation behind the Harman headphone target from the lead researcher himself. http://seanolive.blogspot.com/ I don't necessarily agree 100% with Sean Olives results, but it's more important to understand the underlying thought, better yet from a first hand source.

 

[Robbo99999]

Hmm, I don't think I claimed that they were channel matched. Either way, I certainly didn't intend to do so. What I said (or think I said) was:

They subjectively sounded much more focused to me after applying the EQ curve, which I attributed to corrective channel-matching.

However, after analysing what the EQ actually does, I showed that it does barely anything (0 or 0.1 dB between left and right, close to nothing). With that I tried to say that my initial thought must have been wrong, and that no channel matching happened, or at least not "as immensely" as I experienced it. I then suggested another explanation for this effect, namely that a higher level of (mono) bass leads to a more focused perception.

But all of this is subjective, so if I seemed to implied that my S5X actually is channel matched, or that I could accurately determine that by just listening, then I'm sorry for that, that was unintended.



I am a bit confused about what "the Harman research" actually says about the over-ear curve. In some posts it is described in the way you said (flat speakers in good room, measured in model ears). In some posts it is described as users twisting an EQ until the perception of the headphones was equal to that of the flat speakers in a good room.

I did search ASR and other sites for 'the truth', but there's just sooo soooo much written about all this that I can't find the forest for the trees.




How is a surround field be mixed down to 2 ch? I mean, the mix down in production is clear: you are wearing your Apple Vision and you look in a certain direction, the immersive mixer gives every sound source a HRTF suitable to the location of the source.

But the S5X is a static headphone, no active mixing down. What does it mean in this context to mix a surround field down to 2ch? Has a dummy head been placed in a room where it is hit by audio equally from every direction, and then measure the FR at the dummy eardrum?

Ok, good, I'm glad you didn't say your Ollo was perfectly channel matched, as that wasn't proven.

You're right about the Harman Research, the baseline dummy head measurement was tweaked subjectively in bass & treble by the participants, but the end result is that it would equal the curve of anechoic flat speakers in a room but with a bit more bass & a bit more treble. The 2013 Over Ear Harman Curve actually happens to be equal to anechoic flat speakers in a room, but the 2018 Over Ear Harman Curve has just got a tad more bass & treble vs that one. The 2018 Over Ear Harman Curve is the one used by Amir on this site, as well as the one used in the Oratory EQ's. Therefore when most people talk about Harman Curve for Headphones they're referring to the 2018 version. I have tried listening to the 2013 version, but I normally prefer the 2018 version.

EDIT: for reference, following is the baseline measurement at the dummy head eardrum that was then subsequently tweaked by the participants to add bass & treble, and note that the speakers had first been EQ'd to a flat in-room measurement at listening position, which is not the same thing as just anechoic flat speakers in a room, because flat in-room measurement means no downslope which is actually not desirable, but the researchers wanted to see how the study participants would tweak the bass & treble and if they'd end up close to how anechoic flat speakers measure in a room, which indeed they did. The 2013 Harman Curve is equal to Anechoic Flat speakers in a room, and the 2018 Harman Curve just has a bit more bass & treble vs that. Yes, so following the initial baseline measurement that was then tweaked by the participants for bass & treble:

And then following is a graph showing the various different Harman Targets (for some reason where it says OE 2017, that's actually the OE 2018, the terminology of 2017 vs 2018 seems interchangeable for some reason, but the orange line is the 2018 Harman Target. And IE means in ear, so the green one is a target for IEM's):

 

[Smaestro]

Thank you for that excellent clarification!