ASR Headphone Testing and BK 5128 Hats Measurement System

[Deleted member 16543]

I'm achingly curious to find out whether this option was explored further, either by @amirm or @Mad_Economist or anyone else, for that matter.

Nope

 

[Mad_Economist]

Nope

FWIW, while we've been incommunicado, I've found a three paper series from Sony addressing the role or lack thereof of canal structure on HRTF. My read - and, apparently, B&K's, separately, as they referenced the Sony papers - is that the 4620 canal structure is sufficiently anthropomorphic.

 

[Deleted member 16543]

FWIW, while we've been incommunicado, I've found a three paper series from Sony addressing the role or lack thereof of canal structure on HRTF. My read - and, apparently, B&K's, separately, as they referenced the Sony papers - is that the 4620 canal structure is sufficiently anthropomorphic.

Is that the one used in their 5128 mic?

 

[Mad_Economist]

Is that the one used in their 5128 mic?

Yes, the 4620 is the ear sim (including the canal and pinna) used in the 5128, although apparently you can also buy it standalone.

 

[Deleted member 16543]

Yes, the 4620 is the ear sim (including the canal and pinna) used in the 5128, although apparently you can also buy it standalone.

Well, yeah.. I'd say it's "sufficiently anthropomorphic" too!

 

[Mad_Economist]

Well, yeah.. I'd say it's "sufficiently anthropomorphic" too!

Possibly of interest to you, the third Sony paper includes simulation of sound propagation inside the canal at very high frequencies - I recall that was an interest point.

 

[Deleted member 16543]

Possibly of interest to you, the third Sony paper includes simulation of sound propagation inside the canal at very high frequencies - I recall that was an interest point.

Yes! Thanks

 

[Dennis_FL]

Anyway, any and all feedback is welcome including whether we should even bother doing this.

That's too much money. But as to what to test....the ol' standby - the Sennheiser's legendary 650 of course.

 

[amirm]

That's too much money. But as to what to test....the ol' standby - the Sennheiser's legendary 650 of course.

This fixture is long gone but review of HD650 is already done on the GRAS fixture: https://www.audiosciencereview.com/forum/index.php?threads/sennheiser-hd650-review-headphone.18774/

 

[BYRTT]

That's too much money. But as to what to test....the ol' standby - the Sennheiser's legendary 650 of course.

Excuse me in it sounds like you not up to date in the headphone story, long time ago Amir returned the 5128 analyze system and ended up buy a G.R.A.S 45CA analyze system, legendary 650 Sennheiser's has been reviewed onto both the 5128 & 45CA system, you can find the 650/5128 analyze somewhere into this thread & find the 650/45CA analyze as Amir pointed out in previous post ..

 

[solderdude]

The problem is you will only obtain the difference between the 2 fixtures with the HD650.
When you use a HD800 or a planar and use the same compensation as found using the HD650 the HD800 and planar measured on both fixtures will not be the same.
That correction thus would only apply to HD58X, HD580, HD600 and HD660 (in the same position).
If it were that simple B&K would have already provided a Harman curve for the 5128.

 

[ADU]

There are several different methods that could be used to derive a neutral in-ear response curve for the 5128 rig. Some would be more difficult than others, and require the use of a 5128 rig to perform various measurements. While others would not. And could at least be used to maybe get you in the general ballpark.

 

[Deleted member 16543]

There are several different methods that could be used to derive a neutral in-ear response curve for the 5128 rig. Some would be more difficult than others, and require the use of a 5128 rig to perform various measurements. While others would not. And could at least be used to maybe get you in the general ballpark.

To derive a neutral in-ear response curve for the 5128 rig you need a 5128 rig (also, necessary condition, but not sufficient).
I don't know what methods you have in mind that wouldn't require it. I guess it depends on your definition of ballpark.

 

[amirm]

There are several different methods that could be used to derive a neutral in-ear response curve for the 5128 rig. Some would be more difficult than others, and require the use of a 5128 rig to perform various measurements. While others would not. And could at least be used to maybe get you in the general ballpark.

We developed one based on research that we used in this thread/evaluation. Alas, you have no way of confirming its validity without extensive controlled testing so the results cannot be defended. Indeed what we developed was not correct.

In contrast, I grabbed the target for GRAS45C and ran with it. That unit is one third the price to boot so there is really no reason to use BK 5128.

Also, I find that fitting headphones on the GRAS 45C is much easier than doing the same with B&K 5128. The constraint of a full head with the material that it is built out of makes it harder to make physical corrections. IEMs were next to impossible to stuff in the artificial ears and have them stay there.

 

[Deleted member 16543]

... Indeed what we developed was not correct.

In contrast, I grabbed the target for GRAS45C and ran with it. That unit is one third the price to boot so there is really no reason to use BK 5128.
...

That's a non sequitur. The reason is increased accuracy.. Provided one can develop a correct method.

But I totally understand that if one wants to measure IEMs, the rig shouldn't be that difficult to use.

 

[ADU]

To derive a neutral in-ear response curve for the 5128 rig you need a 5128 rig (also, necessary condition, but not sufficient).
I don't know what methods you have in mind that wouldn't require it. I guess it depends on your definition of ballpark.

Howdy, sax512.

There are several methods that could be used to approximate a neutral response curve without the rig, using the diffuse field response of the rig. Some of these methods would probably be as accurate, if not more so, as the Harman curve on a GRAS rig.

Extrapolating a curve based on the response of several different headphones might also be a possibility. There probably aren't enough headphone measurements currently available for the 5128 to make that practical or reliable at the moment though, I suspect.

If you just want an approximation curve (similar to the Harman target) though, then that should be doable without the rig, using diffuse field calculation imo. The accuracy and reliability of such an approach would probably depend to some extent on how much time and effort you're willing to put into it though.

I can run down a couple of the more easily explainable options along these lines though, if anyone here is still interested. I would also understand though, if some here view this as a can worms that they'd rather not have reopened.... Either way is ok by me.

For the most accurate results though, you would need the rig either for doing some measurements of speakers in a room. And/or to measure a series of different headphones with an approximately neutral response.

 

[ADU]

We developed one based on research that we used in this thread/evaluation. Alas, you have no way of confirming its validity without extensive controlled testing so the results cannot be defended. Indeed what we developed was not correct.

In contrast, I grabbed the target for GRAS45C and ran with it. That unit is one third the price to boot so there is really no reason to use BK 5128.

Also, I find that fitting headphones on the GRAS 45C is much easier than doing the same with B&K 5128. The constraint of a full head with the material that it is built out of makes it harder to make physical corrections. IEMs were next to impossible to stuff in the artificial ears and have them stay there.

Understood Amir. And thank you for the reply.

I haven't had a chance to read through this topic. So I don't know which methods for the 5128 have already been tried here. My offer above with regard to the different methods of deriving a target via diffuse field and so forth were mainly for those who want to try doing some other informal plotting along these lines.

If you'd prefer to have this subject left as is, and not reopened though, I would certainly understand that. And can just let it go. It is totally up to you and the other forum members. I think it's sort of an interesting topic. But I'm ok with whatever the rest of y'all decide.

 

[Deleted member 16543]

I'm for free speech. Intelligible speech would be even better.
I'd suggest to keep your posts short and to the point. It was somewhat difficult to sort through the free form ones in the other thread, to be honest. No offence intended.. it might be just me.
If you could succinctly describe your diffuse field based method I think you will find people willing to give you the benefit of the doubt, here.

 

[ADU]

I'm for free speech. Intelligible speech would be even better.
I'd suggest to keep your posts short and to the point. It was somewhat difficult to sort through the free form ones in the other thread, to be honest. No offence intended.. it might be just me.
If you could succinctly describe your diffuse field based method I think you will find people willing to give you the benefit of the doubt, here.

I can describe them pretty succinctly, sax512. Probably in a just a few lines each (if that much). Provided noone else here has any objections.

 

[amirm]

That's a non sequitur. The reason is increased accuracy.. Provided one can develop a correct method.

Accuracy to what? If I told you I have a new unit of measure for length that is called gzu. I tell you my phone is 3.3 gzu. What does that tell you about the real length of my phone? Nothing because you have no frame of reference for that new measurement.

The unit of measurements we want here is the listener satisfaction. We have that for GRAS. We don't have that for B&K 5128. Furthermore, should one be developed for it, there is no telling currently that it would correlate better with listening test results.

Finally, the claimed accuracy only comes at the highest frequencies where if you breath on the headphone, it changes. For that reason, I more or less ignore everything above 8 to 10 kHz.