Binaural blind comparison test of 4 loudspeakers

[Thomas_A]

Harbeth?

 

[Blumlein 88]

I've no idea really. It's one thing to do a preference test but as to guessing which speaker I'm listening to through binaural recording using an unknown mic setup, of speakers positioned somewhere in an unknown room using recordings, I'm not familiar with.. well let's just say I'd rather look for a needle in a haystack

I agree, I still think the recordings weren't very good as another source of confusion glossing over plenty that makes me not want to guess what speakers these are.

 

[Zvu]

Harbeth?

Oooooooo, good one

 

[phoenixdogfan]

The Bach version sounds a lot like the 1980 Harnoncourt version that I really, really don't like
I didn't vote. I don't particularly like any of the recordings, one was kind of OKish but felt a bit muffled, two sounded bad, 1 was awful to my ears.
@thewas as you noted, too bad we can't choose the tracks.

The worst thing about using a binaural head is you're using a binaural head. Those heads are obviously not using the listener's individual HRTF, which is a significant diminution if not an outright disqualifier to rendering any of those speakers accurately to any individual over headphones.

 

[thewas]

I agree, I still think the recordings weren't very good as another source of confusion glossing over plenty that makes me not want to guess what speakers these are.

Yes, you and few others didn't like them, as you have written repeatedly in this thread. I also don't find them great myself either and also don't find the used music good for such tests. On the other hand the results positively surprised me with their consistency and also matching to what we would expect from the Harman rating theories. And interestingly one of the very few guess till now was even quite good in describing typical loudspeaker tuning classes which all show that they aren't too bad either and/or some people are able to hear some loudspeaker characteristics hrough them.

 

[TimVG]

Yes, you and few others didn't like them, as you have written repeatedly in this thread. I also don't find them great myself either and also don't find the used music good for such tests. On the other hand the results positively surprised me with their consistency and also matching to what we would expect from the Harman rating theories. And interestingly one of the very few guess till now was even quite good in describing typical loudspeaker tuning classes which all show that they aren't too bad either and/or some people are able to hear some loudspeaker characteristics hrough them.

Does that imply there are spinoramas available for each of the models tested?

 

[LTig]

The worst thing about using a binaural head is you're using a binaural head. Those heads are obviously not using the listener's individual HRTF, which is a significant diminution if not an outright disqualifier to rendering any of those speakers accurately to any individual over headphones.

Yes and no. I agree that tonality of (early) binaural recordings (using the Neumann dummy head) sound a bit weird to me, but I have several radio plays where the localisation effects works very good to me (using Sennheiser headphones like my old HD580 without EQ).

 

[thewas]

The worst thing about using a binaural head is you're using a binaural head. Those heads are obviously not using the listener's individual HRTF, which is a significant diminution if not an outright disqualifier to rendering any of those speakers accurately to any individual over headphones.

This of course is the biggest problem of binaural recordings, although it can be partially compensated by EQing individually the listeners headphone to his own HRTF.

But even without it the research has shown surprisingly good correlation to the captured loudpeaker preferences, see https://www.audiosciencereview.com/...st-of-4-loudspeakers.26785/page-6#post-924548

 

[abdo123]

Kii Three
Dutch&Dutch 8C
Kef
B&W

There ya go

C wouldn't be a KEF, not in a million years.

unless it's some shady discontinued model from the previous millennium.

 

[thewas]

Does that imply there are spinoramas available for each of the models tested?

At least some typical on-axis and a couple of axis measurements and for some similar models there are also full spinoramas.

 

[Thomas_A]

This of course is the biggest problem of binaural recordings, although it can be partially compensated by EQing individually the listeners headphone to his own HRTF.

But even without it the research has shown surprisingly good correlation to the captured loudpeaker preferences, see https://www.audiosciencereview.com/...st-of-4-loudspeakers.26785/page-6#post-924548

In addition to that, you would need a binaural curve. No bass boost, and no treble drop. So not even the Etymotic curve would do here.

 

[abdo123]

Actually my 50 cents are on B being a KEF speaker with their weird port tuning (mid-bass hump as a result.

 

[thewas]

In addition to that, you would need a binaural curve. No bass boost, and no treble drop. So not even the Etymotic curve would do here.

Yes and even then a binaural recording and playback is a reduction of a 3D function field/matrix to a single transfer function.

 

[Thomas_A]

Yes and even then a binaural recording and playback is a reduction of a 3D function field/matrix to a single transfer function.

I am soon to record the correct versions with my home-made human setup to compare to the current tracks.

 

[GXAlan]

The worst thing about using a binaural head is you're using a binaural head. Those heads are obviously not using the listener's individual HRTF, which is a significant diminution if not an outright disqualifier to rendering any of those speakers accurately to any individual over headphones.

I agree but what's interesting is that these recordings sound really, really poor in comparison to Binaural recordings from Chesky Records which, even if they aren't ideal from the individual HRTF perspective, are much better and giving you 3D impressions. It's widely available on streaming services. The barbershop track is pretty fun.

Dr. Chesky's Sensational, Fantastic, and Simply Amazing Binaural Sound Show​

 

[thewas]

I agree but what's interesting is that these recordings sound really, really poor in comparison to Binaural recordings from Chesky Records which, even if they aren't ideal from the individual HRTF perspective, are much better and giving you 3D impressions. It's widely available on streaming services. The barbershop track is pretty fun.

Dr. Chesky's Sensational, Fantastic, and Simply Amazing Binaural Sound Show​

There is of course a big difference if you have a binaural recording of a real acoustic event or one of 2 hifi loudspeakers in a typical hifi listening setup and too high listening distance playing a stereo recording, which is also the difference between "live" and typical stereo reproduction.

 

[Zvu]

C wouldn't be a KEF, not in a million years.

unless it's some shady discontinued model from the previous millennium.

I haven't intended to sort them per preference. I just tossed few brands to get the fire going

 

[Thomas_A]

I did not manage to do the recording as good as previously, a bit too low level and noisier. And as before this just microphones taped to the ears pointing at the roof. But at least it is the same. The first file uncorrected and raw and should be listened through binaural headphones, the second one adjusted a bit according to the room signature, and should be listened with Harman curve adjusted headphones (or speakers in normal rooms).

What strikes me is that the true binaural recordings are quite much more detailed, but with a more nasal timbre overall. Setup is probably crucial, but also room signature.

File one, uncorrected.

sing2 raw.wav

Shared with Dropbox

www.dropbox.com

File two, corrected.

sing2 corr.wav

Shared with Dropbox

www.dropbox.com

 

[oivavoi]

Can I haz my loudspeaker reveal before I go to bed please...?

 

[oivavoi]

Ok, here are my guesses before the reveal: i think B is a well-executed standard speaker. I would guess Revel or PSB or one of the good wharfedales, or others. Three-way. A would be the same type, but somewhat inferior (Focal?).

I think D has wider dispersion, but somehow less energy in the mids. I'm not able to say what kind of speaker this would be. Could be some kind of three-way speaker with a high-ish crossover point to the tweeter, like around 5-6 khz, which causes the mid woofer to struggle a bit before the tweeter takes over?

Concerning C - a large two-way speaker, with drivers not aligned optimally? Or something fishy exotic. Single-driver speaker perhaps?

This may be completely wrong though!

EDIT: guess is also based on unreliable memory from listening a couple of days ago