Resolve's B&K 5128 Headphone Target - you can try the EQ's.....

[Robbo99999]

What do you mean with a "10 dB slope"? Linear descending slope, ok, but 10 dB between which frequencies? I usually see this referred like -0.8 dB/octave, so the octave (double of frequency) specified the interval, but yours does not. Thanks in advance.

I think they mean they're applying a -10dB step change EQ Function across the whole frequency range from 20Hz to 20kHz - so they're taking the Diffuse Field Curve and then applying a linear slope to that which has 0dB of change at 20Hz, and -10dB of change at 20kHz.....so it's a linear slope change of -10dB applied (-10dB across that whole audible range). You could work out what that would equate to in terms of -XdB per octave, but I've not calculated what that would be as an equivalent, Resolve could probably tell you straight off the bat.

EDIT: there's 10 octaves within the range from 20Hz-20000Hz, so Resolve's -10dB slope is the equivalent of -1dB per octave. (I can' remember if they're including a bass shelf within that, I don't think they are.)

 

[Resolve]

To me it seems the 8kHz peak isn't of the same "ilk" as the 9kHz cancellation dip that naturally occurs in GRAS. That 9kHz cancellation dip in GRAS is large sharp and unpredictable in nature

This is one where I'd defer to @oratory1990. From what I've tested, headphones behave similarly for these two features, with similar unpredictability. But it could also be that the DF HRTF should reflect that. We never cared about this feature on GRAS, but maybe with an unsmoothed target its worth including.

 

[Robbo99999]

This is one where I'd defer to @oratory1990. From what I've tested, headphones behave similarly for these two features, with similar unpredictability. But it could also be that the DF HRTF should reflect that. We never cared about this feature on GRAS, but maybe with an unsmoothed target its worth including.

Well, that's my interpretation, I'm just suggesting you could explore testing whether that 8kHz feature really matters or not, with my one idea being to listen to sine sweeps in that area after the headphones had been EQ'd to include that 8kHz feature (to check if they sound flat in that area) - but of course that would require that you would deviate from the diffuse field approach, so that speaks to redefining the initial approach. That 8kHz peak is a relic of that quick & dirty Harman work translating from GRAS to B&K, and yet you see it in the HD600 too, it's just food for thought.

 

[Resolve]

with my one idea being to listen to sine sweeps in that area after the headphones had been EQ'd to include that 8kHz feature (to check if they sound flat in that area) - but of course that would require that you would deviate from the diffuse field approach

It seems not all headphones behave the same way for that region, on both systems. So even if there's a trend, there's still an argument not to include it in the target. But yeah I've seen it reflected in Sean's stop-gap method. Either way though it doesn't prevent the use of the DFHRTF approach - at least I don't quite see how that's incompatible. Just have to identify the effects of the canal entrance resonance. At the moment I'm of the opinion it shouldn't be included, just like it isn't on the GRAS DFHRTF and it isn't with Harman, but if there's a compelling reason to include it then I see no issue with doing that.

 

[Robbo99999]

It seems not all headphones behave the same way for that region, on both systems. So even if there's a trend, there's still an argument not to include it in the target. But yeah I've seen it reflected in Sean's stop-gap method. Either way though it doesn't prevent the use of the DFHRTF approach - at least I don't quite see how that's incompatible. Just have to identify the effects of the canal entrance resonance. At the moment I'm of the opinion it shouldn't be included, just like it isn't on the GRAS DFHRTF and it isn't with Harman, but if there's a compelling reason to include it then I see no issue with doing that.

I suppose if you decide to not include that 8kHz feature then you'll also have to remember to not EQ headphones down to the target curve in that area, which is a bit of a tricky headphone by headphone decision, because they'll be times when a headphone does have real excess energy in that area that would ideally need to be EQ'd out. If it was me I'd do two different Target Curves, one being your current Diffuse Field Adjusted Target and the other one perhaps Harman's quick & dirty Target Curve (that has the 8kHz peak feature), then I'd measure and EQ all the models of headphones I had access to, I'd EQ them to each of the two targets and listen to slow sine sweeps through that controversial 8kHz area to see if it sounds like it is flat through that area, making a note of if it sounds flat or a dip or a peak there. (I might also take off the headphone and reseat it on my head & run the test again to ensure it wasn't just due to placement variation). After doing that for each of the headphones you might find that there is a pattern that one of the two Target Curves provides the most flat sounding sine sweep in that 8kHz area - which would give you guidance on whether you need to revisit the creation of the Target Curve or use one of the two you tested (in & around that 8kHz region anyway). You might need to do slow sine sweeps using known Anechoic Flat Speakers through that 8kHz area just to check your natural hearing is flat through that area before you start listening to sine sweeps on headphones, because for example my hearing has a bit of a quirk where I have a large hearing boost between 8-12kHz, so I wouldn't EQ down peaks I hear in headphones in that area because that's just a natural part of my own hearing. I've never embarked on listening to sine sweep project on this scale so I don't know how feasible this would be and if you will come across pitfalls, but I have on occasion listened to slow sine sweeps on my my various headphones to verify if my headphones had the same peaks that were measured by Oratory on his GRAS for instance - a lot of the times there's parity between the two with the peaks being in the same place, and sometimes you'll hear a few peaks that aren't in the measurements, but generally I've found there's good parity between measured peaks on GRAS vs listening to slow sine sweeps on my headphones in those areas. (Wide peaks I find hard to identify as the rate of change is slow, but narrow steeper peaks I find a lot more easy to identify when listening to sine sweeps.) Following is the website I used, using the arrow keys to cycle up & down the frequencies:

Online Tone Generator - generate pure tones of any frequency

www.szynalski.com

P.S. Listening to sine tones is pretty unpleasant & sounds like ringing in the ears, so I would suggest only doing one headphone per day and trying to get through the testing session as quickly as possible, because I find your ears eventually try to tune out the ringing sounds of the sine tone so I think you become desensitised to the test if you do it for too long, and then you can't pick up on the fine variations. It's not really a pleasant process, lol! (Just do like one headphone a day and get them all done over a couple weeks or something.)

 

[isostasy]

Does anyone know about the B&K 4128? DMS' latest video uses that to measure the HE-1 and now I'm just getting confused. I can't find it used for any other headphone measurements but can see a mention of 'IEC 60318-4' in the spec sheet so is it meeting the same standard as the GRAS KEMAR and therefore comparable?

 

[oratory1990]

Does anyone know about the B&K 4128? DMS' latest video uses that to measure the HE-1 and now I'm just getting confused. I can't find it used for any other headphone measurements but can see a mention of 'IEC 60318-4' in the spec sheet so is it meeting the same standard as the GRAS KEMAR and therefore comparable?

The 4128C would be described in 60318-5 (dash 5, not dash 4).
But yes, it is generally comparable to the version of the Gras Kemar that we are talking about, as in it has rubber pinnae and an ear canal simulator based on the 711 ear simulator.
The pinna is not identical to the ones used on Kemar though, so you get different results above ~3 kHz.
It's used quite a lot in acoustic engineering, just as the Kemar is.

 

[markanini]

I have compensated this error in my stereo speakers via mid/side EQ before I discovered the pano phase shuffler. Mid/side EQ might also be the way to go with IEMs/headphones instead of baking it into the target curve.

(emphasis mine)

Yes.

# Mid-Side EQ
Copy: MID=R+L SIDE=R+-1.0*L
Channel: MID
#... filters on mid
# https://www.audiosciencereview.com/forum/index.php?threads/monophonic-vs-stereophonic-timbre-change.15829/post-508549
# Stereo Center Image fix
Filter: ON PK Fc 1850 Hz Gain 2 dB Q 2.87
Filter: ON PK Fc 2800 Hz Gain -2 dB Q 2.87
Filter: ON PK Fc 7200 Hz Gain -2 dB Q 4.32
Channel: SIDE
#... filters on side
Copy: R=0.5*MID+0.5*SIDE L=0.5*MID+-0.5*SIDE

This is my code that used. But I moved to pano phase shuffler since.

Have you used this kind of EQ with any IEM or headphone of yours? Did you enjoy the results? I wanna try this out.

 

[digitalfrost]

Have you used this kind of EQ with any IEM or headphone of yours? Did you enjoy the results? I wanna try this out.

Haven't tried it with headphones yet, it didn't occur to me to re-create an error on purpose . If you just invert the decibel values you can start. You can see I didn't fully compensate the values from the graph and the graph is for a normally reflective room. Since we don't have that with headphones I guess increasing the values towards full compensation would make sense. Maybe try

Filter: ON PK Fc 1850 Hz Gain -4 dB Q 2.87
Filter: ON PK Fc 2800 Hz Gain 3 dB Q 2.87
Filter: ON PK Fc 7200 Hz Gain 3 dB Q 4.32

 

[Merkurio]

Leaving aside the more intricate technical aspects of the discussion and the supposed benefits of the B&K 5128, I must say that this Resolve's target sounds terrible to me compared to the old-school Harman target from a GRAS system (directly from Oratory presets or cautiously curated using AutoEQ) in every headphone I tested, even those with allegedly low variability between units.

Bright in the upper-mid region, not enough bass and overall a very artificial presentation.

Not my thing, for sure.

 

[Robbo99999]

Leaving aside the more intricate technical aspects of the discussion and the supposed benefits of the B&K 5128, I must say that this Resolve's target sounds terrible to me compared to the old-school Harman target from a GRAS system (directly from Oratory presets or cautiously curated using AutoEQ) in every headphone I tested, even those with allegedly low variability between units.

Bright in the upper-mid region, not enough bass and overall a very artificial presentation.

Not my thing, for sure.

I think they've decided to increase the downward slope that they apply to their Diffuse Field from 0.8dB per octave to 1dB per octave since they first put out those EQ's, so that will increase the bass also as well as removing some brightness. They're in the process of reworking their target and probably working out the next best steps to take in the project, so it should sound different to their EQ's that are currently at the thread I linked in the first post.

 

[Resolve]

Leaving aside the more intricate technical aspects of the discussion and the supposed benefits of the B&K 5128, I must say that this Resolve's target sounds terrible to me compared to the old-school Harman target from a GRAS system (directly from Oratory presets or cautiously curated using AutoEQ) in every headphone I tested, even those with allegedly low variability between units.

Bright in the upper-mid region, not enough bass and overall a very artificial presentation.

Not my thing, for sure.

Yeah once again, what you're listening to is NOT my target. If you read through the thread, the entire purpose is to have users indicate if there's too much ear gain in the upper mids at 3khz. Please comment with your adjustment at 3khz. Did you reduce it by 2dB? 3dB? We want to know if adding a point of adjustment there makes sense, or if we can get by simply by tilting the DFHRTF more.

 

[GaryH]

this really isn't "our target"

Yes, it is. No-one else with a 5128 uses a default target consisting of such a fine-grained DFHRTF with a linear slope and not shelf. Not Sean Olive. Not Sam Vafei formerly of Rtings now Linus Tech Tips. Not Soundguys. Not Jude of Head-Fi (well he doesn't use a target at all, so as not to upset all his sponsors and pluralist Head-Fier subjectivists). Not even your buddy / business partner Crinacle does. And you have created the specific EQs to this target that you're asking people's opinion of. Now a lot of people are saying they don't really like it, it's too bright, too shouty, all those similar descriptive words (sound familiar?), and you say 'it's not our target'...

When it comes to its fine-grained features, this is literally just the Diffuse Field HRTF. If people take issue with that, it entails taking issue with the Diffuse Field HRTF

Too right people take issue with that. Acoustic scientists like Sean Olive for example:

Several studies done by Lorho (2009), Fraunhofer (2012) and Harman (2013-2019) have since shown there are more preferred targets than DF. Instead of measuring the loudspeaker in a DF they are measured in a semi-reflective listening room as a starting point. Why? Because most recordings are not optimized for playback in reverberation chambers because most listeners don't listen in them but rather semi-reflective rooms with an average RT60 of 0. 4 s. There is a strong frontal direct sound component and some strong early reflection contributions from lateral and other directions -- not at all like a DF field where the arrivals at the listener are random and uniformly distributed in direction. The rooms tend to reinforce the bass below 200 Hz so the in-room response is not flat.

Since recordings are optimized through loudspeakers in a semi-reflective room, they sound best over headphones that emulate that response. And there are lots of experimental data support it.

And your justification for 'slope not shelf' is also flawed. The test conducted by Harman in which the former was preferred was done using an Audeze LCD2, which is an acoustic outlier in that it has an almost completely airtight front volume (so much so people have complained about a suction effect when taking them off, which I believe Audeze rectified by introducing some front venting in other later models). This all comes back to your old friend bass slam/impact. As I've told you before, all else equal perceived bass impact likely correlates with degree of front volume seal. A bass shelf predominantly accentuates low and sub-bass, which also correlates with bass impact. When you combine the two effects, some/many are likely to find this high bass impact too overpowering, and so with these (very particular) headphones prefer a slope, which does not predominantly emphasize the low/sub-bass. This would also explain why DF was not rated that badly with these headphones, as they naturally provide some bass impact even with flat DF-bass. The other test in this same study (which funnily enough you rarely if ever mention) used the Sennheiser HD518, with a much more common non-airtight front volume design, with which shelf was rated considerably higher than DF and with a higher rating than both on the LCD2. Again, the difference in results between the headphones can be explained by front seal differences, with the bass shelf providing the perceived bass impact needed without an airtight front volume, the latter few (if any modern?) headphones have, and so test results involving such outlier headphones should not be used to base targets around.

One thing I will add, we've moved to using a 10dB slope

Yeah once again, what you're listening to is NOT my target. If you read through the thread, the entire purpose is to have users indicate if there's too much ear gain in the upper mids at 3khz. Please comment with your adjustment at 3khz. Did you reduce it by 2dB? 3dB? We want to know if adding a point of adjustment there makes sense, or if we can get by simply by tilting the DFHRTF more.

When you (headphones.com) have to make so many attempts at fixes to your hypothetical target that's supposed selling points were that it was 'more based in theory', and 'more fine-grained', fixes like smoothing out the 3 kHz peak which go directly against the latter, and are theoretically arbitrary and post-hoc and so go against the former; it's time to accept that your whole approach is just fundamentally flawed. And the reasons why are laid out above.

 

[amirm]

Folks both don't learn and think this kind of stuff is easy. When many people complain about the heavily researched Harman curve, how on earth does a person wake up in the morning thinking they can do one better without said research? I know I learned my lesson when I was using the 5128.....

 

[Merkurio]

Yeah once again, what you're listening to is NOT my target. If you read through the thread, the entire purpose is to have users indicate if there's too much ear gain in the upper mids at 3khz. Please comment with your adjustment at 3khz. Did you reduce it by 2dB? 3dB? We want to know if adding a point of adjustment there makes sense, or if we can get by simply by tilting the DFHRTF more.

Sorry if I got confused with the target name, I'm not very up to date on it (beyond your video) and I definitely tried experimenting by increasing the amount of bass a bit and reducing that extra spice around 3 kHz, but it still doesn't match the results I've gotten with other GRAS measurements, either through Oratory presets or by making my own adjustments using AutoEQ with the Harman target as a base, which is how I get the best results (usually with Crinacle and Squig's EQ tools).

Maybe you're right and DF target is not my thing.

The test conducted by Harman in which the former was preferred was done using an Audeze LCD2, which is an acoustic outlier in that it has an almost completely airtight front volume (so much so people have complained about a suction effect when taking them off, which I believe Audeze rectified by introducing some front venting in other later models). This all comes back to your old friend bass slam/impact. As I've told you before, all else equal perceived bass impact likely correlates with degree of front volume seal. A bass shelf predominantly accentuates low and sub-bass, which also correlates with bass impact. When you combine the two effects, some/many are likely to find this high bass impact too overpowering, and so with these (very particular) headphones prefer a slope, which does not predominantly emphasize the low/sub-bass.

Interesting, this seems to explain why the LCD2Cs had such a "punchy" bass response compared to other fully open headphones I tried in the past, despite measuring virtually flat on the graphs...

Folks both don't learn and think this kind of stuff is easy. When many people complain about the heavily researched Harman curve, how on earth does a person wake up in the morning thinking they can do one better without said research? I know I learned my lesson when I was using the 5128.....

I, from ignorance, don't quite understand the benefits of using this B&K measurement rig and everyone turning to it suddenly.

I've the idea that it will present a FR closer to what we hear (never the same, because of our personal HRTF issue), but anyway, those benefits are present after 9-10 kHz, an area whose accuracy may not serve us too much to make better EQ decisions anyway, no?

So, what's the point? For R&D and production environments I could understand its added precision, or to evaluate more accurately the behavior of a headphone by looking at the graph (particularly >10 kHz) before making a purchase decision, but honestly I don't quite get how it benefits us from an equalization point of view.

What is Dr. Sean Olive's position on this?

 

[Robbo99999]

Yes, it is. No-one else with a 5128 uses a default target consisting of such a fine-grained DFHRTF with a linear slope and not shelf. Not Sean Olive. Not Sam Vafei formerly of Rtings now Linus Tech Tips. Not Soundguys. Not Jude of Head-Fi (well he doesn't use a target at all, so as not to upset all his sponsors and pluralist Head-Fier subjectivists). Not even your buddy / business partner Crinacle does. And you have created the specific EQs to this target that you're asking people's opinion of. Now a lot of people are saying they don't really like it, it's too bright, too shouty, all those similar descriptive words (sound familiar?), and you say 'it's not our target'...

Too right people take issue with that. Acoustic scientists like Sean Olive for example:

And your justification for 'slope not shelf' is also flawed. The test conducted by Harman in which the former was preferred was done using an Audeze LCD2, which is an acoustic outlier in that it has an almost completely airtight front volume (so much so people have complained about a suction effect when taking them off, which I believe Audeze rectified by introducing some front venting in other later models). This all comes back to your old friend bass slam/impact. As I've told you before, all else equal perceived bass impact likely correlates with degree of front volume seal. A bass shelf predominantly accentuates low and sub-bass, which also correlates with bass impact. When you combine the two effects, some/many are likely to find this high bass impact too overpowering, and so with these (very particular) headphones prefer a slope, which does not predominantly emphasize the low/sub-bass. This would also explain why DF was not rated that badly with these headphones, as they naturally provide some bass impact even with flat DF-bass. The other test in this same study (which funnily enough you rarely if ever mention) used the Sennheiser HD518, with a much more common non-airtight front volume design, with which shelf was rated considerably higher than DF and with a higher rating than both on the LCD2. Again, the difference in results between the headphones can be explained by front seal differences, with the bass shelf providing the perceived bass impact needed without an airtight front volume, the latter few (if any modern?) headphones have, and so test results involving such outlier headphones should not be used to base targets around.


When you (headphones.com) have to make so many attempts at fixes to your hypothetical target that's supposed selling points were that it was 'more based in theory', and 'more fine-grained', fixes like smoothing out the 3 kHz peak which go directly against the latter, and are theoretically arbitrary and post-hoc and so go against the former; it's time to accept that your whole approach is just fundamentally flawed. And the reasons why are laid out above.

Some good points GaryH, but I don't totally agree with your sentiment in the last paragraph, I don't mind if Resolve & Co experiment with new target curves on the 5128. Regarding Resolve saying it's not his target is semantics and I think he was a bit wrong (& unnecessary) to say that in reply to Merkurio, but Merkurio has since then posted that he did indeed try experimenting with lowering the 3kHz peak as per Resolve's initial instructions & intention, so he gave it a good shot, so that's all good.

 

[amirm]

I've the idea that it will present a FR closer to what we hear (never the same, because of our personal HRTF issue), but anyway, those benefits are present after 9-10 kHz, an area whose accuracy may not serve us too much to make better EQ decisions anyway, no?

Exactly.

So, what's the point? For R&D and production environments I could understand its added precision, or to evaluate more accurately the behavior of a headphone by looking at the graph (particularly >10 kHz) before making a purchase decision, but honestly I don't quite get how it benefits us from an equalization point of view.

That is what B&K told me as to its main application and not for reviewing headphones.

 

[Merkurio]

Exactly.

That is what B&K told me as to its main application and not for reviewing headphones.

Well, this seems to be more detrimental than positive from a hobbyist perspective.

We are starting to see even more clone measurement equipment based on GRAS pinnas and IEC711 318-4 couplers directly from Aliexpress, with an accuracy at least "comparable" to the original systems at some extent, which opens up a world of possibilities for those that we're interested in entering the world of measurements personally, either because we are headphone nerds or because we seek maximum precision with our personal units to develop a better EQ, and now suddenly reviewers decide to change the measurement paradigm (one well established and fuctional) by another without clear or even existing benefits.

What can I tell you, it seems a bit hasty and disappointing to me.

 

[amirm]

What can I tell you, it seems a bit hasty and disappointing to me.

My sentiments exactly as I stated at the start and throughout the thread. It is as if the headphones.com had money to burn and folks decided to do the burning for them! You pay more and get less and confuse the market as a bonus.

 

[Merkurio]

Doing a critique exercise, I think all of this it's a byproduct of the average reviewer mindset (and I'm not saying that as a personal attack on Resolve, or that it's inherently a bad thing), it's just their minds are constantly focused looking for new and cutting edge information about the audio industry, and that's fine, but the mistake in my opinion is to believe that everything is determined by a "higher spec" and lack common sense in the most elementary questions.

I.E.: If something that plays a critical role in a specific function isn't used and endorsed by experts in audio research yet (such as Sean Olive and the Harman team), then don't try to reinvent the wheel and pretend that you know more than the rest of them, especially when you're not an expert on the subject!