• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Binaural blind comparison test of 4 loudspeakers - II

Which speaker comes closest to the original recording?

  • Speaker A

    Votes: 7 25.9%
  • Speaker B

    Votes: 4 14.8%
  • Speaker C

    Votes: 3 11.1%
  • Speaker D

    Votes: 13 48.1%

  • Total voters
    27
  • Poll closed .

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,425
Likes
7,941
Location
Brussels, Belgium
So who is next on setting up a binaural speaker challenge? Very entertaining subject with both suprises and hopefully some learning. :)
I hope the next one would be without speakers with obvious flaws.

a comparision between Something like Neumann KH80 and Genelecs 8341A .etc would be really interesting for example. So far these tests just show 'bad speakers sound bad'. It's entertaining don't get me wrong, but it's not really the epiphany we're looking for imo.
 
Last edited:

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,721
Likes
1,935
Location
Oslo, Norway
The best sounding loudspeaker in this test is apparently GRIMM audio LS1 . The only loudspeaker in the test that has dsp crossover and driven active. I´m not surprised at all. A good active design will always sound better, even compared with the best passive loudspeaker.

I agree with you. I didn't write this previously, but it's actually true and not just hindsight - when I listened to the speakers I thought that several of them seemed a bit off in what I very subjectively would call "timing". I noticed this mostly on the piano track, but also on track nr 3. In some ways this is absurd, of course, as no loudspeakers lose time or play slower than the others, but it could be about phase etc in the crossovers. Loudspeaker D sounded most "right" and coherent in its overall presentation to me, as I think I wrote. In the Grimm LS1 white paper they write a bit about doing crossovers which doesn't mess up the phase, which becomes much easier through DSP. I do believe in the active advantage, even when we keep frequency response etc constant.

EDIT: but I was way off in some of my other subjective assessments, so these subjective listening impressions don't say that much in the grand scheme of things! My own hearing is no guide to audio reality, of course
 

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,721
Likes
1,935
Location
Oslo, Norway
I hope the next one would be without speakers with obvious flaws.

a comparision between Something like Neumann KH80 and Genelecs 8340A .etc would be really interesting for example. So far these tests just show 'bad speakers sound bad'. It's entertaining don't get me wrong, but it's not really the epiphany we're looking for imo.

Would you really classify Grimm LS1 as a "bad speaker"? I would find such a claim a bit surprising...

The KEF in this test has a dispersion pattern which is not that different from some of the Neumann and Genelec models, I think
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,425
Likes
7,941
Location
Brussels, Belgium
Would you really classify Grimm LS1 as a "bad speaker"? I would find such a claim a bit surprising...
No. but all the other speakers have obvious flaws. the KEF Reference 3 might look good at an instance but it's not tuned flat. Basically all the other speakers have obvious flaws thus the Grimm wins the comparison. There is no new information here.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,423
Likes
2,408
Location
Sweden
It could be that one or the other does not like the results. Therefore, right from the start, there is no conspiracy and the data were not manipulated by me.

The result of the poll:
View attachment 160089


Speaker A - Heco Direct (3000€)
View attachment 160090 View attachment 160094


Speaker B - KEF Reference 3 (9000€)
View attachment 160091 View attachment 160095

Speaker C - Bowers & Wilkins 802 D3 (22000€) (No manufacturer link found)
View attachment 160092View attachment 160096

Speaker D - Grimm Audio LS1 (22000€) with subwoofer modul
View attachment 160093 View attachment 160100

Source of the frequency response measurements is hifitest.de.
Source of the binaural recordings is German magazin lowbeats soundcloud.com database..

The current binaural recordings (by lowbeats.de) are made in an EBU-Tech 3276 compliant studio. The older recordings, used in our poll, were made in their own acoustically treated listening room (here is the link to the translation or use the translator of your choice). In the pictures you can see the typical setup.
View attachment 160102
Two MBHO small-diaphragm condenser microphones with free-field equalized omnidirectional characteristics in AB configuration were used for the binaural recording.
The current binaural recordings are performed with a KU100.
More details about their binaural recordings can be found here (here is the link to the translation or use the translator of your choice).

Is there a measured steady state curve for this room at LP?
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,525
Likes
37,058
No. but all the other speakers have obvious flaws. the KEF Reference 3 might look good at an instance but it's not tuned flat. Basically all the other speakers have obvious flaws thus the Grimm wins the comparison. There is no new information here.
The new info is letting us compare speakers recorded this way. Makes good sense to start with something having obvious differences. Maybe next something with 4 outstanding speakers would be better.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,423
Likes
2,408
Location
Sweden
There are still questions from me regarding how these files should be listened to, headphone-wise. This picture was from another thread regarding the Etymotic Target curve and a link from Sean Olive:


If I understand this, you have both an impact in bass and from 1 kHz and up depending if you have EQ the room response to flat or not. So to correct a "Harman curve optimised" headphone response to "flat", you must increase the HF region from 1 kHz and gradually up to 10-20 kHz by 4-5 dB, and decrease the bass by 4-5 dB, which I also did with my files in the thread "binaural session 1". This made it sound much more like it sounds in the room.

So this leaves me the question how the room response really is in the current room where the binaural recording was done. Is it "flat" or does it follow the sloping Harman target, given a speaker like the Revel F208 with flat on-axis and smooth off-axis playing in that room?
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,763
Likes
16,239
I hope the next one would be without speakers with obvious flaws.

a comparision between Something like Neumann KH80 and Genelecs 8340A
The KH80 and 8340 without obvious flaws? Now that is far fetched, maybe a 8351b but not less.
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,763
Likes
16,239
Please demonstrate the flaws in the 8341A spinorama please.
You wrote above 8340 and not 8341...
8341 is similar to my above suggestion of its bigger brother 8351b, just more deep bass and SPL limited.
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,763
Likes
16,239
@thewas what do you have against the KH80 btw? Seems like another flawless design.
Very bass and spl limited, meaning also high imd at not very high listening levels, it's a tiny 2 way in the end. Also vertical directivity has the usual lobe like all non coaxials but just not too bad due to the small drivers and high order crossover.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,425
Likes
7,941
Location
Brussels, Belgium
Very bass and spl limited, meaning also high imd at not very high listening levels, it's a tiny 2 way in the end. Also vertical directivity has the usual lobe like all non coaxials but just not too bad due to the small drivers and high order crossover.

I agree but that doesn’t mean that the design is flawed, it’s very difficult to match the drivers so well like they did. It might be the most linear non-coaxial speaker in this forum.
 

Ron Texas

Master Contributor
Joined
Jun 10, 2018
Messages
6,084
Likes
8,918
A very interesting topic. Too bad the Grimm is rather expensive. The Heco Direct reminds me of the old Snell EIII which is now made in a fancy version by Audio Note.
 
OP
C

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,088
Location
.de, DE, DEU
So this leaves me the question how the room response really is in the current room where the binaural recording was done. Is it "flat" or does it follow the sloping Harman target, given a speaker like the Revel F208 with flat on-axis and smooth off-axis playing in that room?

In post#73 there is the link about the used listening room for the binaural recordings.
Above 200Hz the reverberation time of the listening room is 0.25 seconds and in the bass range 0.6 seconds. That is pretty good for a room with an original area of 70m².

Thus, the in-room response of an anechoic flat-measuring speaker should be evenly sloping.

The slope of the in-room frequency response should perhaps even be somewhat stronger than the PIR from the Spinorama predicts, since the PIR applies to "average" listening rooms, but the listening room in which the binaural recordings were made is acoustically heavily treated.
.
 

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
A very interesting topic. Too bad the Grimm is rather expensive. The Heco Direct reminds me of the old Snell EIII which is now made in a fancy version by Audio Note.
An interesting thing for the DIYer is that GRIMM is very open how and why they have done things. The drivers used can be bought.
A skilled DIY:er can make a clone very easy.
 

Honken

Senior Member
Joined
Mar 10, 2020
Messages
339
Likes
602
Location
Scania
Maybe the results make the Heco a steal at € 1600 a pair (current price)
Can you really find them at that price locally? Here they're quite a bit more expensive (~3000€) but you can find the Einklang for about 1600€. Looking at it for a millisecond, I assume that this is a wildly different design though.

Still though, even for the ~3000€ that the Direkt goes for locally - it's dirt cheap compared to the competition!
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,423
Likes
2,408
Location
Sweden
In post#73 there is the link about the used listening room for the binaural recordings.
Above 200Hz the reverberation time of the listening room is 0.25 seconds and in the bass range 0.6 seconds. That is pretty good for a room with an original area of 70m².

Thus, the in-room response of an anechoic flat-measuring speaker should be evenly sloping.

The slope of the in-room frequency response should perhaps even be somewhat stronger than the PIR from the Spinorama predicts, since the PIR applies to "average" listening rooms, but the listening room in which the binaural recordings were made is acoustically heavily treated.
.
In that case, you would need a headphone that has a target curve according to the green line in the Olive link. Otherwise I don't understand how you can get the correct impression of the tonality of the recordings. A "Harman target" headphone will sound dark and bass-heavy. The files however looks as if there is a general peaking at 10 kHz by 4-5 dB in all recordings.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,425
Likes
7,941
Location
Brussels, Belgium
In that case, you would need a headphone that has a target curve according to the green line in the Olive link. Otherwise I don't understand how you can get the correct impression of the tonality of the recordings. A "Harman target" headphone will sound dark and bass-heavy. The files however looks as if there is a general peaking at 10 kHz by 4-5 dB in all recordings.

if you check the spectra graphs that @ctrl made the deviations are barely enough to make the dramatic effect that you're mentioning. the direct sound ofcourse is still flat after all.
 
Top Bottom