• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

KEF R Series with MAT white paper

OP
davidbosch

davidbosch

Member
Audio Company
Joined
Mar 15, 2021
Messages
41
Likes
413
Location
United Kingdom
Has anyone already compared the old R Series vs the R Meta? The graphs in Kef's white papers look pretty similar to me and prices for the old R series are getting really low.

I've seen the old R3 for 1000-1200€ and the R5 for 1500-2000€.
R3 Meta is 2200€ (+700€ for the stand) and R5 Meta 3500€.

Not sure if this rather large price difference is justified.
Hi @juliangst , the responses in the 2018 R Series white paper are not Spinoramas, this is before we decided to use Spinoramas to present the speaker's response, so they're not entirely comparable to the Spinoramas on the R Meta white paper. In the Crossover section in the R Meta white paper (https://audiosciencereview.com/forum/index.php?threads/kef-r-series-with-mat-white-paper.43946/) we show a comparison of the on axis, power average and horizontal and vertical averages for R3 2018 vs R3 Meta. The rest of the document presents the several improvements in the Meta models, particularly the Uni-Q, which is considerably more advanced. You can also look at the measured THD figures for the tweeter, midrange and system (R11), which show pretty clear improvements.

I do agree that at discounted prices, 2018 R Series right now is a steal, particularly when you see what you can get in the market for that money, at least retail.
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,299
Likes
14,809
Hi @juliangst , the responses in the 2018 R Series white paper are not Spinoramas, this is before we decided to use Spinoramas to present the speaker's response, so they're not entirely comparable to the Spinoramas on the R Meta white paper. In the Crossover section in the R Meta white paper (https://audiosciencereview.com/forum/index.php?threads/kef-r-series-with-mat-white-paper.43946/) we show a comparison of the on axis, power average and horizontal and vertical averages for R3 2018 vs R3 Meta. The rest of the document presents the several improvements in the Meta models, particularly the Uni-Q, which is considerably more advanced. You can also look at the measured THD figures for the tweeter, midrange and system (R11), which show pretty clear improvements.

I do agree that at discounted prices, 2018 R Series right now is a steal, particularly when you see what you can get in the market for that money, at least retail.
First of all a big thank you from my side for contributing so helpful here, really appreciated.

When looking at the comparison you mention

1682495510633.png

it seems the total directivity indexes are almost identical which is probably a result of the not really changed geometry, crossover frequency and slopes(?) so my question is if someone would equalise the old R3 to match the responses of the R3 Meta closely, they would probably sound tonaly very similar, only distortion, max SPL etc would differ audibly, what do you think?

I own and really still enjoy daily both the original LS50 (with full bandwidth EQ and subwoofer, now my desktop system) and LS50 Meta (with subwoofer and EQ only in the modal region, my main system) and with equalisation the old LS50 can be brought closer to the LS50 Meta but not as close as above R3 vs R3 Meta because the two LS50 generations differ more in the directivity.
 

tifune

Major Contributor
Joined
Jan 18, 2020
Messages
1,057
Likes
743
Hi @tifune, I think I rather not, as I'm not posting as a civilian here, so apologies. Controlled directivity into low frequency sounds ideal, but in a room below the Schroeder frequency I'm not sure it's yet well understood how such speaker will interact with the room modes.

That makes sense; appreciate your candor. Is there any reading you could point me toward specifically around the topic?
 

Mario Sanchez

Active Member
Joined
Jan 11, 2021
Messages
161
Likes
275
Can a little bit more information be provided on the crossover topologies? Seems like a 2nd order one between the uni-q midrange and the tweeter, but can't be sure...
 

Opal

Member
Joined
Jul 4, 2022
Messages
49
Likes
127
I haven't caught up with all the conversations however I would say that what one listens to is the whole system, and the driver(s) is only part of the story. The enclosure inside and outside, its built and materials, the crossover, the final balance etc. all contribute to the end result. To say the R3 Meta should sound equal to the Ref 1 non-Meta kind of disregards that the Ref 1 non-Meta still has a better LF driver, a better cabinet (and larger), better crossover, etc. The Uni-Q is from 2014, so in that area the R3 Meta wins: low distortion MF motor, better tweeter, tweeter gap damper, metamaterial absorber on the tweeter, smaller MF suspension, flexible decoupling chassis, stiffening ribs on the tangerine rear side, etc. I think this doesn't align with what you experienced listening to R3 meta vs Ref 5 non-Meta, but again, the end result could be more than the Uni-Q.
Hi David, May i ask is the Ref Meta Uni-Q same as the R Meta Uni-Q? If not, is the Ref Meta's Uni-Q superior in many ways? The price difference is...... :oops:
 

sifi36

Active Member
Joined
Aug 7, 2021
Messages
118
Likes
237
Thank you @davidbosch for being so open and willing to answer questions.

One thing I’ve always wondered about with Reference versus R series is about tolerances. Are pair matching and driver performance tolerances the same for both?
 

Jukebox

Active Member
Joined
Apr 11, 2020
Messages
192
Likes
346
Thank you @davidbosch for being so open and willing to answer questions.

One thing I’ve always wondered about with Reference versus R series is about tolerances. Are pair matching and driver performance tolerances the same for both?
R is made in China where they probably have a factory standard unit to match for an entire batch of speakers.
Reference is made in England where every speaker and crossover is measured against the factory standard unit; and you also get a paper with the actually frequency response measurement.
It will only be normal to think that Reference is the safe bet here, as matching and tolerances goes.
But I guess the R line holds their ground very well!

1682601451248.png
1682601535490.png
 
Last edited:

stren

Active Member
Forum Donor
Joined
Jan 9, 2020
Messages
243
Likes
287
Thank you @davidbosch for being so open and willing to answer questions.

One thing I’ve always wondered about with Reference versus R series is about tolerances. Are pair matching and driver performance tolerances the same for both?

My understanding was:
- Uni Q is similar - but visually I thought the reference always have more ribs so I'm assuming different and possibly better. I heard something about better metals used, maybe in the magnets, but not sure.
Also
- Better crossovers (is this different order or just better quality components yielding to higher accuracy of frequency?)
- Reference has bigger, heavier, better designed cabinets yielding less resonance, distortion, port noise (higher material costs and shipping)
- Bass drivers are better, better magnets (maybe?) and presumably more xmax and linearity
- Better shadowflare (one piece front plate, maybe smoother?)
- Tighter testing, specs, control
- Built(or assembled) in UK vs China
- More expensive finishes
 

juliangst

Addicted to Fun and Learning
Joined
Dec 11, 2021
Messages
898
Likes
889
Is it just me or do reference meta speakers in black look way darker in real life than they do on the renders?

On most advertisement pictures the baffle has a dark grey industrial metal look to it (which I really like visually) but on most real life pics they just look pitch black (especially in contrast to the drivers)
 

sifi36

Active Member
Joined
Aug 7, 2021
Messages
118
Likes
237
R is made in China where they probably have a factory standard unit to match for an entire batch of speakers.
Reference is made in England where every speaker and crossover is measured against the factory standard unit; and you also get a paper with the actually frequency response measurement.
It will only be normal to think that Reference is the safe bet here, as matching and tolerances goes.
But I guess the R line holds their ground very well!

I’m aware of this, just curious to know how similar or different the tolerances are between their Chinese vs UK manufactured speakers. One assumes that the drivers for the UK assembled speakers are still made in China. That being said, are the tolerances for pair matching the same or not?
 

exm

Active Member
Joined
Sep 21, 2021
Messages
223
Likes
191
My understanding was:
- Uni Q is similar - but visually I thought the reference always have more ribs so I'm assuming different and possibly better. I heard something about better metals used, maybe in the magnets, but not sure.
Also
- Better crossovers (is this different order or just better quality components yielding to higher accuracy of frequency?)
- Reference has bigger, heavier, better designed cabinets yielding less resonance, distortion, port noise (higher material costs and shipping)
- Bass drivers are better, better magnets (maybe?) and presumably more xmax and linearity
- Better shadowflare (one piece front plate, maybe smoother?)
- Tighter testing, specs, control
- Built(or assembled) in UK vs China
- More expensive finishes

I'll let you David answer this one but I believe that the Uni-Q drivers are different between the R and Reference Series. Basically, the Reference Uni-Q is of a higher quality.
 

fineMen

Major Contributor
Joined
Oct 31, 2021
Messages
1,504
Likes
666
... it seems the total directivity indexes are almost identical ... someone would equalise the old R3 to match the responses of the R3 Meta closely ...
Me? I frankly tell you, the now totally outdated sub-par old R3s are good enough for engineers who anyway won't believe in true sound. The speaker's personality is strong enough to stand the dire situation playing for a measurement microphone alone, I hope.

Statistically the preferrence score refers to somebody elses verdict.

So why don't you try? The LS50 experienced a hype for its cute looks, maybe. The R-series was something else and still is. Is it really wothwhile to discuss the breed's pedigree being Chinese or British, counting (or not) its cold metallic ribs? Explore your preferrence. The Reference being a safe bet in the Meta empire, so I was told here, if not a bit costly. But who cares?
 

Descartes

Major Contributor
Joined
Dec 27, 2020
Messages
1,712
Likes
823
Hand made always cost more bespoke suit will fit better than the one on the racks!

Now moving that analogy to electronics makes me smile, otherwise Apple would have jumped on this to make more money!
 

pjn

Active Member
Forum Donor
Joined
Aug 28, 2021
Messages
161
Likes
143
Me? I frankly tell you, the now totally outdated sub-par old R3s are good enough for engineers who anyway won't believe in true sound. The speaker's personality is strong enough to stand the dire situation playing for a measurement microphone alone, I hope.

Statistically the preferrence score refers to somebody elses verdict.

So why don't you try? The LS50 experienced a hype for its cute looks, maybe. The R-series was something else and still is. Is it really wothwhile to discuss the breed's pedigree being Chinese or British, counting (or not) its cold metallic ribs? Explore your preferrence. The Reference being a safe bet in the Meta empire, so I was told here, if not a bit costly. But who cares?
I was seriously considering Reference 1s rather than R7s but it seemed to me that (1) they were designed for a larger room than mine (15ft x 15 x 8) and (2) as a result I would probably not get much better performance than the R7s as I wouldn't be able to optimally place them without still having problems with room modes (or making them worse!). Again, for me, one of my LS50s had serious resonance problems at 90Hz which forced me to cross them over at >130Hz, so actually worse technical performance than my venerable 104/2s, although LS50s did a better job over about 800 Hz
 

stren

Active Member
Forum Donor
Joined
Jan 9, 2020
Messages
243
Likes
287
I know this is r thread but seeing as the Kef guys are here... Any plans for a centre version of the ls60w and/or adding wysa support? The thinness of the ls60w is really useful for slotting under a TV horizontally. I'm guessing the market is too small to justify but I can dream right?

Also the old walnut r11's are on clearance at best buy for less than half the price of the r11 metas.

Clearance R11 non meta vs R7 meta vs Used Reference 1 Non meta vs Used LS60... all in that 3 to 6K range...
 
Last edited:
OP
davidbosch

davidbosch

Member
Audio Company
Joined
Mar 15, 2021
Messages
41
Likes
413
Location
United Kingdom
First of all a big thank you from my side for contributing so helpful here, really appreciated.

When looking at the comparison you mention

View attachment 281519
it seems the total directivity indexes are almost identical which is probably a result of the not really changed geometry, crossover frequency and slopes(?) so my question is if someone would equalise the old R3 to match the responses of the R3 Meta closely, they would probably sound tonaly very similar, only distortion, max SPL etc would differ audibly, what do you think?

I own and really still enjoy daily both the original LS50 (with full bandwidth EQ and subwoofer, now my desktop system) and LS50 Meta (with subwoofer and EQ only in the modal region, my main system) and with equalisation the old LS50 can be brought closer to the LS50 Meta but not as close as above R3 vs R3 Meta because the two LS50 generations differ more in the directivity.
The DIs do look almost the same. The 2018 speakers have a pretty decent MF-HF integration but there was room for improvement. The crossover is a bit different, the parallel capacitor in the HF circuit is deleted, so the filter is a second-order high pass whereas in the 2018 one the parallel circuit functioned as a notch. The crossover frequency is also lower (2.3 vs 2.9 kHz). This helped improve the off axis response clearly within +/-45 deg.

If you take the 2018 R3 and EQ it to have a similar measured response to the R3 Meta, I think you won't be awfully far, but I don't think they will really sound the same. The MF and HF drivers are quite different in distortion and other behaviour. The Meta also has much improved behaviour in at least 7 resonances, and in my experience, since these can sound quite annoying when you hit them while listening to music, one tends to balance the speaker slightly to attenuate the frequency regions where they happen. As you point out, distortion will also be different, generally higher in 2018 R3. Even at normal volume, harmonic distortion will play a part in how the speaker sounds to us, and thus have an effect on the balance we end up with. Then small things, like the capacitor in series with the tweeter, have (for some reasons we understand and some other we don't fully yet) a considerable effect on how the speaker sounds with music, which can't really be seen in the frequency response.

If you can get a really cheap 2018 R3, EQ or not, it's already a bargain of a speaker. EQ'ing it closer to the R3 Meta might make it better, might not be enough probably. I would just go further and EQ it like you would an active speaker to really try to smooth the response and then I think that would be more interesting.
 
OP
davidbosch

davidbosch

Member
Audio Company
Joined
Mar 15, 2021
Messages
41
Likes
413
Location
United Kingdom
Can a little bit more information be provided on the crossover topologies? Seems like a 2nd order one between the uni-q midrange and the tweeter, but can't be sure...
Hi Mario, if you have a look at the white paper, there's a figure with the crossover topology. The HF high pass is second order electrical, the MF low pass is first order electrical. The acoustical transfer functions of the filtered drivers are of higher order though thanks to their natural roll off and sum correctly around crossover.
 
OP
davidbosch

davidbosch

Member
Audio Company
Joined
Mar 15, 2021
Messages
41
Likes
413
Location
United Kingdom
Hi David, May i ask is the Ref Meta Uni-Q same as the R Meta Uni-Q? If not, is the Ref Meta's Uni-Q superior in many ways? The price difference is...... :oops:
Hi @Opal, they're not the same. You can have a look at the Blade and Reference Meta white paper which explains that Uni-Q in detail. There's several features we trickled down from that to the R Meta Uni-Q though.
 
OP
davidbosch

davidbosch

Member
Audio Company
Joined
Mar 15, 2021
Messages
41
Likes
413
Location
United Kingdom
Thank you @davidbosch for being so open and willing to answer questions.

One thing I’ve always wondered about with Reference versus R series is about tolerances. Are pair matching and driver performance tolerances the same for both?
Hi @sifi36, Blade and Reference have pair-matched Uni-Qs and measurement tolerances are particularly very tight. R Series are obviously manufactured with a line standard as well and are of great quality but they don't have pair-matched Uni-Qs. Part of the Blade and Reference offer is that they are hand-assembled by a technician and they do take many hours to complete and pass their tests.
 
OP
davidbosch

davidbosch

Member
Audio Company
Joined
Mar 15, 2021
Messages
41
Likes
413
Location
United Kingdom
My understanding was:
- Uni Q is similar - but visually I thought the reference always have more ribs so I'm assuming different and possibly better. I heard something about better metals used, maybe in the magnets, but not sure.
Also
- Better crossovers (is this different order or just better quality components yielding to higher accuracy of frequency?)
- Reference has bigger, heavier, better designed cabinets yielding less resonance, distortion, port noise (higher material costs and shipping)
- Bass drivers are better, better magnets (maybe?) and presumably more xmax and linearity
- Better shadowflare (one piece front plate, maybe smoother?)
- Tighter testing, specs, control
- Built(or assembled) in UK vs China
- More expensive finishes
Hi @stren. The cone ribs are a visual cue only. The Uni-Q on Blade/Reference Meta is of higher performance, the MF motor uses Neodymium, but there's several other differences, you can check them out in the white paper. All your other points are correct.
 
Top Bottom