Binaural blind comparison test of 4 loudspeakers - II

[thewas]

Maybe the results make the Heco a steal at € 1600 a pair (current price)

My exact thoughts, beautiful design, acoustic advantages of the wide baffle, good drivers (Heco/Magnat uses Klippel tools and software to optimise them) and measurements and seemingly a nice tuning for a really good price.

Some measurements:

(at 95 dB)
source: https://www.hifitest.de/test/lautsprecher_stereo/heco-direkt_12446

source: https://www.stereo.de/hifi-test/produkt/heco-direkt-1098

source: https://www.i-fidelity.net/testberichte/high-end/heco-direkt/labor-seite-6.html

 

[Blumlein 88]

Seems like speaker B should have done better maybe into 2nd place looking at the measurements. Any idea why it seemed to be picked 3rd by almost everyone commenting? We can see why the B&W did poorly.

 

[RHO]

Seems like speaker B should have done better maybe into 2nd place looking at the measurements. Any idea why it seemed to be picked 3rd by almost everyone commenting? We can see why the B&W did poorly.

Too much, boomy, bass?

 

[thewas]

Too much, boomy, bass?

Exactly, as I had commented before the results were revealed:

...
Speaker B doesn't sound bad either, has the most low bass which is a bit too much without room correction.
...
It wasn't an easy choice, I just decided on A based on the third song (D being very close), with room EQ I might have taken B instead.

 

[PierreV]

Seems like speaker B should have done better maybe into 2nd place looking at the measurements. Any idea why it seemed to be picked 3rd by almost everyone commenting? We can see why the B&W did poorly.

The effect the binaural recording technique has on our perception of reflected sound that you described in post 52? The downslope which is supposed to be the preferred response (and which is my preferred response in real life for most music) may be emphasized.
In general, I find that I prefer the more direct stereo recording that "thealphaaudio" used in their grimm vs atc vs d&d video (and probably the revel vs focal linked above). It comes much closer to my real-life perception than binaural recordings.

 

[Thomas_A]

The effect the binaural recording technique has on our perception of reflected sound that you described in post 52? The downslope which is supposed to be the preferred response (and which is my preferred response in real life for most music) may be emphasized.
In general, I find that I prefer the more direct stereo recording that "thealphaaudio" used in their grimm vs atc vs d&d video (and probably the revel vs focal linked above). It comes much closer to my real-life perception than binaural recordings.

So one interesting analysis would be the in-room spectrum vs the original. My plan is to make my own recording and some analysis but have no time until Wednesday or Thursday. I will post results when I have done that.

 

[oivavoi]

Concerning the external validity of this test, I tend to think that the results may attest to a probable "real" preference for these speakers, in terms of ranking - even though reproducing speakers over binaural recordings through headphones doesn't sound like real speakers in a room. It makes sense that the LS1 are preferred by most, and that the B&W is least preferred. Given the fact that none of the speakers are disastrous when it comes to frequency response, it also makes sense that many of us perceived the test as challenging. EDIT: I also think that the LS1 may have scored better due to a relative smoothness in the overall presentation which may be related to the fact that it has active crossovers.

As to the relative lack of preference for the KEF, it may of course have to do with the bass. I honestly can't remember for sure how I perceived them, so I'll not second guess my own reaction. If I re-listen now it would be non-blind thus more biased. It may have to do with the bass response in the room, of course. But can we rule out that it's related to how the waveguide affects the sound? I don't think we can rule that out. After all it's the speaker with the deepest waveguide of the four.

Would also be interesting if anybody who knows how to do it could do a spectral analysis of the sound from the speakers as recorded by the microphone, compared to the original recordings.

 

[ctrl]

Once again the best-measuring loudspeaker seems to be preferred by the majority. Cool.

Yep!
The KEF loudspeaker probably has the best directivity, but due to the on-axis frequency response and LW, which drops significantly to high frequencies, it can't play out this superiority.
Good directivity alone does not make a very good speaker, it depends on the crossover fine tuning.

I find it extremely interesting that the 802D3 received three votes. If the question had been "Which loudspeaker do you like best", I would not have been surprised, but in direct comparison with the original recording ("Which comes comes closest to the original recording?"), the 802D3 obviously deviates tonally the most. For these three voters, tonality does not seem to have been the top priority, but other acoustic parameters - fascinating.

That Heco Direkt is actually an interesting speaker, at least with some eq. Haven't thought too much about that speaker previously. Looks good (IMO), decent measurements, and now I know that I don't perceive it to be that far behind the LS1, which I have perceived as absolutely in sighted in-person listening...

This speaker surprised me the most. For a short time, it looked like it would be a neck-and-neck race with the LS1.
An old-school design with a large woofer and wide baffle, and due to the high sensitivity of 91-92dB, somewhat "thin" in the low frequency response.
It is also interesting that no one has described the loudspeaker with 91-92 dB sensitivity as having a particularly dynamic sound behavior, as is often automatically assumed.

Seems like speaker B should have done better maybe into 2nd place looking at the measurements.

With this model from KEF, the frequency response of the listening window is very likely to drop almost 5dB in the 500Hz to 10kHz range - that might just be a bit too much.

 

[abdo123]

looks like neither directivity or dynamic ability made a difference.

 

[ctrl]

So one interesting analysis would be the in-room spectrum vs the original.

First a comparison of the In-room frequency response of speaker A to D

 

[pma]

Seems like speaker B should have done better maybe into 2nd place looking at the measurements. Any idea why it seemed to be picked 3rd by almost everyone commenting? We can see why the B&W did poorly.

I am quite sure that in case of speakers, measurements still do not tell enough and listening evaluation is a must. Too many variables including room acoustics and listeners taste.

 

[Thomas_A]

First a comparison of the In-room frequency response of speaker A to D
View attachment 160130

Could you show one speaker at the time vs the original also?

 

[ctrl]

The following diagrams show the in-room spectra of the loudspeakers in comparison to the spectrum of the original recording.

All speaker spectrum are normalized to the spectrum of the original recording.
In the interpretation it is crucial that it is not decisive how close a loudspeaker spectrum comes to the original (0dB line), but that it runs as evenly as possible parallel to it.
The 0dB line was roughly set by me. You can mentally shift the frequency response curves of the speakers up or down as you like.

A) Heco Direct


B) KEF Reference 3


C) Bowers & Wilkins 802 D3


D) Grimm Audio LS1

 

[RHO]

The following diagrams show the in-room spectra of the loudspeakers in comparison to the spectrum of the original recording.

All speaker spectrum are normalized to the spectrum of the original recording.
In the interpretation it is crucial that it is not decisive how close a loudspeaker spectrum comes to the original (0dB line), but that it runs as evenly as possible parallel to it.
The 0dB line was roughly set by me. You can mentally shift the frequency response curves of the speakers up or down as you like.

A) Heco Direct
View attachment 160131

B) KEF Reference 3
View attachment 160132

C) Bowers & Wilkins 802 D3
View attachment 160133

D) Grimm Audio LS1
View attachment 160134

Microphone roll-off above 8.5kHz?

 

[oivavoi]

The following diagrams show the in-room spectra of the loudspeakers in comparison to the spectrum of the original recording.

All speaker spectrum are normalized to the spectrum of the original recording.
In the interpretation it is crucial that it is not decisive how close a loudspeaker spectrum comes to the original (0dB line), but that it runs as evenly as possible parallel to it.
The 0dB line was roughly set by me. You can mentally shift the frequency response curves of the speakers up or down as you like.

A) Heco Direct
View attachment 160131

B) KEF Reference 3
View attachment 160132

C) Bowers & Wilkins 802 D3
View attachment 160133

D) Grimm Audio LS1
View attachment 160134

Thanks. Very interesting. This was not what I thought I heard. I'm a little confused on how to interpret this. Or rather: I understand what the graphs are telling me, but these graphs don't make it intuitively clear to me why most people preferred the LS1, with the Heco in second place, and why the Kef and B&W came out last

 

[Vuki]

Hearing is believing in this case.

 

[Thomas_A]

Clearly the KEF lost on its peaky bass. Otherwise it is the speaker that has a bit more energy at 2 kHz vs 3-4 kHz. Which may be one reason that I liked the piano from KEF more than from the others.

 

[thewas]

C) Bowers & Wilkins 802 D3

Thank you again, interesting to see how the this way room averaged deviations in the mids don't look as nasty as they sound, except maybe the large upper bass hump around 150 Hz, which is probably also related from us perceiving direct and reflected sounds differently while in such a "listening position kind of sound power" spectrum they get all mixed together.

 

[abdo123]

The following diagrams show the in-room spectra of the loudspeakers in comparison to the spectrum of the original recording.

All speaker spectrum are normalized to the spectrum of the original recording.
In the interpretation it is crucial that it is not decisive how close a loudspeaker spectrum comes to the original (0dB line), but that it runs as evenly as possible parallel to it.
The 0dB line was roughly set by me. You can mentally shift the frequency response curves of the speakers up or down as you like.

A) Heco Direct
View attachment 160131

B) KEF Reference 3
View attachment 160132

C) Bowers & Wilkins 802 D3
View attachment 160133

D) Grimm Audio LS1
View attachment 160134

I still don't understand why would the Grimm Audio measure in reality better when the KEF has the better spin?

 

[ctrl]

This was not what I thought I heard. I'm a little confused on how to interpret this. Or rather: I understand what the graphs are telling me, but these graphs don't make it intuitively clear to me why most people preferred the LS1, with the Heco in second place, and why the Kef and B&W came out last

Would suspect that "real" IN-room measurements would provide a bit more information.
You can associate the sound impressions with the spectra a little bit.

The bloated bass of the KEF can be easily spotted.
The overall very even response of the GrimmAudio speaker is noticeable.
The early low bass drop of the Heco, compared to the B&W or Grimm, is noticeable.
The B&W seems to be out of line, but if you look at the response in the 500-5000Hz range, this speaker shows the greatest "ripples" there.