• 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!

KEF Reference 1 META Bookshelf Speaker Review by Erin's Audio Corner

phoenixdogfan

Major Contributor
Forum Donor
Joined
Nov 6, 2018
Messages
3,294
Likes
5,068
Location
Nashville
I do see that the measured on axis responce is slightly flatter for this speaker vs the R3.
It would be interesting to A/B/Q them.

Perceived brightness is dependent on so many factors. Think about High Passing a favorite pair of speakers at 120hrz, then add the bass back, then add a robust sub.

Now a dynamic speaker on an intense passage vs a less dynamic one. A speaker with slightly elevated treble on a sweep, might sound less bright due to presenting less dynamic SPL.

Dispersion is also huge factor for me in perceived brightness. Low distortion is as well.
One thing we hardly ever see IMD testing of tweeters. Some tweets sing at high SPL in complex areas and others are really stressed even if HD on sweep is fine.

Interesting for me at least, my pair of R3's did not really measure with that slightly elevated treble when simultaneously measured with several other speakers.

In practice, in my case the R3's never once sounded siblant, brash, harsh, shrill or anything along that line(unless it was due to poor recordings) and I listen to some crazy stuff sometimes.

Who knows.
That is the subjective part.
Over-all both speakers measure so well.

Shame I didn't dig the R3's. Did a lot of testing/listening, appreciate the speaker but it just wasn't quite there for me.

The Reference 1's are to $ for me to try currently. Maybe someday.
You can do comparisons like Ref 3 vs Meta Reference 1 over on Pierre Aubert's site. Both speakers are there, just select compare drop down box, and select each speaker. It will give on axis, listening window, early reflections, lateral and sound power directivity--the whole suite of 2034 measurements in one graph.


 

bo_knows

Addicted to Fun and Learning
Joined
Aug 17, 2020
Messages
797
Likes
788
Location
Dallas, Texas USA
Does any one know if the meta uni-q driver is the same in the reference and lower tier meta models?

I would like to know if this new driver can be ordered from KEF as a part. If any one knows part number for this driver, please share.

The distortion and compression profile looks great. If the price of the drivers is not ridiculous, I entertain the idea of swapping the drivers on my R11s.
Even if they (KEF) sold you the meta uni-q driver, it will not be a direct swap. They will also need to provide you with the updated crossover hence why they are not offering this option to the previous Blade and reference owners.
 

gags11

Senior Member
Forum Donor
Joined
Feb 8, 2019
Messages
356
Likes
528
You can do comparisons like Ref 3 vs Meta Reference 1 over on Pierre Aubert's site. Both speakers are there, just select compare drop down box, and select each speaker. It will give on axis, listening window, early reflections, lateral and sound power directivity--the whole suite of 2034 measurements in one graph.



Thank you for the link, this is awesome!

I’m mostly intrigued with the new meta driver by it’s ability to play high SPL levels without compressing. Erin’s reviews show appreciable compression with R3s that is absent in the new reference meta. I couldn’t see compression data on the site, unless I’m missing something.
 

gags11

Senior Member
Forum Donor
Joined
Feb 8, 2019
Messages
356
Likes
528
Even if they (KEF) sold you the meta uni-q driver, it will not be a direct swap. They will also need to provide you with the updated crossover hence why they are not offering this option to the previous Blade and reference owners.

My understanding is that the difference is in execution and way to keep distortion to minumum. I don’t expect significant difference in crossover design.

I may be wrong in this, as I have always thought that crossovers can be source of distortion and compression.
 
OP
thewas

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,739
Likes
16,161
Even if they (KEF) sold you the meta uni-q driver, it will not be a direct swap. They will also need to provide you with the updated crossover hence why they are not offering this option to the previous Blade and reference owners.
Correct, also in the past they did sell spare drivers only with proof of loudspeaker ownership, which is unfortunate for many DIY people would love to have them.
 

ROOSKIE

Major Contributor
Joined
Feb 27, 2020
Messages
1,915
Likes
3,394
Location
Minneapolis
You can do comparisons like Ref 3 vs Meta Reference 1 over on Pierre Aubert's site. Both speakers are there, just select compare drop down box, and select each speaker. It will give on axis, listening window, early reflections, lateral and sound power directivity--the whole suite of 2034 measurements in one graph.


Thanks and I do know of thar site that but I am speaking of doing an ABX listening test.
It would be fun to ABX the R3 and Reference 1 along with other speakers of course.
 

gags11

Senior Member
Forum Donor
Joined
Feb 8, 2019
Messages
356
Likes
528
Correct, also in the past they did sell spare drivers only with proof of loudspeaker ownership, which is unfortunate for many DIY people would love to have them.

So be it, my KEF R11s will keep their inferior driver! …which beats almost anything out there. I’m more happy if KEF is not selling this for DIY. After all, a few fraction of a dB of compression at certain frequencies is not worth thousands of dollars,
 

Vacceo

Major Contributor
Joined
Mar 9, 2022
Messages
2,630
Likes
2,749
Correct, also in the past they did sell spare drivers only with proof of loudspeaker ownership, which is unfortunate for many DIY people would love to have them.
Gutting out my IQ's for an improved driver and crossover. Sounds quite good, actually. :D
 

phoenixdogfan

Major Contributor
Forum Donor
Joined
Nov 6, 2018
Messages
3,294
Likes
5,068
Location
Nashville
Thank you for the link, this is awesome!

I’m mostly intrigued with the new meta driver by it’s ability to play high SPL levels without compressing. Erin’s reviews show appreciable compression with R3s that is absent in the new reference meta. I couldn’t see compression data on the site, unless I’m missing something.
No, the limitation the Harman score and the Klippel measurements it's based on, is the failure to include the power handling/dynamic range dimension of a speaker in the assessment. That's how something like a Neuman KH 80 can have a score comparable to a Revel Salon 2 (6.2 vs 6.3).
 
Last edited:

ROOSKIE

Major Contributor
Joined
Feb 27, 2020
Messages
1,915
Likes
3,394
Location
Minneapolis
I’m mostly intrigued with the new meta driver by it’s ability to play high SPL levels without compressing. Erin’s reviews show appreciable compression with R3s that is absent in the new reference meta. I couldn’t see compression data on the site, unless I’m missing something.
Hi, where are you seeing the data on the EAC compression test for the R3? I did not see the results of the test published nor in the video. I must have missed it somehow.

In my testing I do not believe the R3 is suffering from such issues as long as it is High Passed where the midbass unit runs out of steam. It can play very loudly, very cleanly. Point to where me where the compression simulation test is if you are able, I'd like to check it out.

Speaking of the 3 sine wave sweep that Erin uses. I do appreciate the test and it is helpful, I would caution anyone trying to make a 1-1 association for the actual loudspeakers compression qualities with music/content.
A sweep is not dynamic like music and it is not a whole mother load of frequencies being played simultaneously like music. Additionally with music/content you may be playing loudly for extended periods of time and the TS parameters of the drivers will change and so therefore will the frequency response of the system. In some cases it may not be compression being exposed in the sense of a driver hitting limits but rather a new frequency response due to essentially a new environment.
You also have varying pressures in the box so one can see how that could really complicate things especially when the source is not a sweep but again a bunch of frequencies all at once and changing in SPL very dynamically.
There may also be issues with the amplifier in the test reacting to different loads and at different SPL's. I don't known what amp he is using in that regard.
In any case very appreciated measurements and personally I would hesitate to draw a firm conclusion from them most of the time.
Dr. Earl Geddes and others at the time argued that edge diffraction was still a problem on speakers with large waveguides. For example, here is a picture of a GedLee Summa by @Patrick Bateman. To give you a sense of scale, that's a 15" woofer, and a Danley PA horn (I think model SH50) next it. As you can see there's very little flat area between the WG termination and the cabinet edge roundover.

IMG_0661.JPG


I'm not convinced the little extra flare is the whole answer. I suspect the difference is more computing power to model cabinet effects and design around them by playing with relative dimensions, features, and even the crossover. Obviously KEF's waveguide stub came from that, but more likely than not so did the cabinet dimensions for late model Revels and so on as well. Keep in mind that the current JBL waveguides (M2, 4367, 7-series, SCL) are, as I understand them, diffraction managing shapes rather than minimum diffraction shapes such as the GedLee oblate-spheroid waveguide.
Another interesting thing portion of Geddes published work is basically summed up in the quote from some PP slides.
This was based on data gathered from blind listening tests.
This relates to both Edge Diffraction and modest amounts of non-linear distortion such as compression or frequency response deviations in the system as driver parameters and pressures in the enclosures are in flux when in use.
1651978909662.png

1651978945470.png

1651978993333.png

1651979067640.png
 

jonfitch

Senior Member
Joined
Sep 2, 2019
Messages
479
Likes
532
Hi, where are you seeing the data on the EAC compression test for the R3? I did not see the results of the test published nor in the video. I must have missed it somehow.

In my testing I do not believe the R3 is suffering from such issues as long as it is High Passed where the midbass unit runs out of steam. It can play very loudly, very cleanly. Point to where me where the compression simulation test is if you are able, I'd like to check it out.

Speaking of the 3 sine wave sweep that Erin uses. I do appreciate the test and it is helpful, I would caution anyone trying to make a 1-1 association for the actual loudspeakers compression qualities with music/content.
A sweep is not dynamic like music and it is not a whole mother load of frequencies being played simultaneously like music. Additionally with music/content you may be playing loudly for extended periods of time and the TS parameters of the drivers will change and so therefore will the frequency response of the system. In some cases it may not be compression being exposed in the sense of a driver hitting limits but rather a new frequency response due to essentially a new environment.
You also have varying pressures in the box so one can see how that could really complicate things especially when the source is not a sweep but again a bunch of frequencies all at once and changing in SPL very dynamically.
There may also be issues with the amplifier in the test reacting to different loads and at different SPL's. I don't known what amp he is using in that regard.
In any case very appreciated measurements and personally I would hesitate to draw a firm conclusion from them most of the time.

Another interesting thing portion of Geddes published work is basically summed up in the quote from some PP slides.
This was based on data gathered from blind listening tests.
This relates to both Edge Diffraction and modest amounts of non-linear distortion such as compression or frequency response deviations in the system as driver parameters and pressures in the enclosures are in flux when in use.
View attachment 205314
View attachment 205315
View attachment 205316
View attachment 205317

Are these effects even audible at normal listening levels? It would be useful if they described at what decibel levels these would be an issue.
 

ROOSKIE

Major Contributor
Joined
Feb 27, 2020
Messages
1,915
Likes
3,394
Location
Minneapolis
No, the limitation the Harman score and the Klippel measurements it's based on, is the failure to include the power handling/dynamic range dimension of a speaker in the assessment. That's how something like a Neuman KH 80 can have a score comparable to a Revel Salon 2 (6.2 vs 6.3).
Well the Harman score is not based on any Klippel measurements.
It was a collaboration of blind listening tests and data from anechoic measurements.

I say this as the Klippel is different from what Harman used and therefore it is possible that the results could vary. It is even possible the Klippel is higher resolution. In any case Harman did not use a Klippel.

One is used here or course.

I do think the score is a bit dubious. I like it some days and others NSM. Toole and Olive did make an appearance here to poo-poo the score a bit and that was cool. It ought to be noted : They do have a vested interest now(or @least Olive does as an employee of Hraman) in reducing emphasis on the score due to competitors reaching higher scores. That said I do believe that they were being honest enough in stating Harman is not using the score as a goal.

Along with not involving the power handling, dynamic capabilities of the speaker it also ignores some others things that may be important including a big one, loudspeaker dispersion.(especially in stereo, perhaps less of a factor when surrounded by 5,10 or 15 speakers)
Also you have to factor in how the loudspeaker behaves in a room in the bass region where SBIR is at play. A speaker with one 6" bass driver behaves differently vs one with 3, stacked 6" bass drivers even if they have similar output capabilities.
It also is not calculated differently for speakers that have been intentionally designed for off axis listening. It is calculated on axis each time.
 
Last edited:

ROOSKIE

Major Contributor
Joined
Feb 27, 2020
Messages
1,915
Likes
3,394
Location
Minneapolis
Are these effects even audible at normal listening levels? It would be useful if they described at what decibel levels these would be an issue.
Sorry I meant to add a link

The maximum average SPL was only 80db due to being conducted in a clinical setting. It was meaningfully visible in the data to Geddes in testing from the low 70's to 80db at the ear. One can imagine these effects increase even more as SPL drives toward rocking out, orchestra at home and home theater warfare.
 
Last edited:

gags11

Senior Member
Forum Donor
Joined
Feb 8, 2019
Messages
356
Likes
528
Hi, where are you seeing the data on the EAC compression test for the R3? I did not see the results of the test published nor in the video. I must have missed it somehow.

In my testing I do not believe the R3 is suffering from such issues as long as it is High Passed where the midbass unit runs out of steam. It can play very loudly, very cleanly. Point to where me where the compression simulation test is if you are able, I'd like to check it out.

Speaking of the 3 sine wave sweep that Erin uses. I do appreciate the test and it is helpful, I would caution anyone trying to make a 1-1 association for the actual loudspeakers compression qualities with music/content.
A sweep is not dynamic like music and it is not a whole mother load of frequencies being played simultaneously like music. Additionally with music/content you may be playing loudly for extended periods of time and the TS parameters of the drivers will change and so therefore will the frequency response of the system. In some cases it may not be compression being exposed in the sense of a driver hitting limits but rather a new frequency response due to essentially a new environment.
You also have varying pressures in the box so one can see how that could really complicate things especially when the source is not a sweep but again a bunch of frequencies all at once and changing in SPL very dynamically.
There may also be issues with the amplifier in the test reacting to different loads and at different SPL's. I don't known what amp he is using in that regard.
In any case very appreciated measurements and personally I would hesitate to draw a firm conclusion from them most of the time.

Another interesting thing portion of Geddes published work is basically summed up in the quote from some PP slides.
This was based on data gathered from blind listening tests.
This relates to both Edge Diffraction and modest amounts of non-linear distortion such as compression or frequency response deviations in the system as driver parameters and pressures in the enclosures are in flux when in use.
View attachment 205314
View attachment 205315
View attachment 205316
View attachment 205317

Here are Erin’s results for the latest reference and R5 results. The R5 has smaller woofers, but the midrange driver is the same.
 

Attachments

  • D5CCC4F4-F852-4FA7-A930-90049E496D83.png
    D5CCC4F4-F852-4FA7-A930-90049E496D83.png
    86.4 KB · Views: 142
  • 31B99C7B-8BA9-411A-A791-09C34105A461.png
    31B99C7B-8BA9-411A-A791-09C34105A461.png
    77.5 KB · Views: 139

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
Is there a real difference between the Uni-Q drivers in R3 or R1 Meta ? Is it the same driver ?
 

Mnyb

Major Contributor
Forum Donor
Joined
Aug 14, 2019
Messages
2,636
Likes
3,593
Location
Sweden, Västerås
It’s not the same driver the new reference series seems to share uni q with the new blade series ? Afaik ? Seems developed together . Everyone is quite curious about if and when a “meta” uni q appears in the R series and the redesign of xover and other adjustments and what it would bring ? We have a whole tread about that :)
 

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
It’s not the same driver the new reference series seems to share uni q with the new blade series ? Afaik ? Seems developed together . Everyone is quite curious about if and when a “meta” uni q appears in the R series and the redesign of xover and other adjustments and what it would bring ? We have a whole tread about that :)
Yes - the difference in frequency response between R3 and R1 Meta might be entirely dependent on different crossover designs . Ím sceptical to the suggestion that there are 2-3 different uni-Q drivers out there , that looks the same. It should be much cheaper for the company to create only one optimal version and then let the crossover make the sound differences between different series, and the rest is pure marketing. Maybe Im wrong about this , but we really need to see real measurements of the drivers to see if they really are different.
 

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
Maybe Im wrong about this , but we really need to see real measurements of the drivers to see if they really are different.

Yep. You’re wrong. ;)

Somewhere around 2013 I measured the raw drivers from the Q100, R300/500 and the LS50. The data is on my site. The differences scale up to current models.

They’re all unique with different purposes. The surround alone is a hint.
 

Tangband

Major Contributor
Joined
Sep 3, 2019
Messages
2,994
Likes
2,789
Location
Sweden
Yep. You’re wrong. ;)

Somewhere around 2013 I measured the raw drivers from the Q100, R300/500 and the LS50. The data is on my site. The differences scale up to current models.

They’re all unique with different purposes. The surround alone is a hint.
Thanks for this info . :)
 
Top Bottom