• 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). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

KEF R3 Speaker Review

tecnogadget

Active Member
Joined
May 21, 2018
Messages
254
Likes
436
Location
Madrid, Spain
Thanks for the info :)
Oh wow I did not know they would implement a mechanical low pass like that. Pretty bad compromise if you ask me, the driver must have a poor Qms and not be very responsive.
That much increase in 2-3k area is odd. Making the paralel with filter, its a low pass filter with very high Q (increase of energy before cutoff, followed by a sharp drop after cutoff.).
Of course more damping means better frequency response, but that damping happens between voice coil and cone, so the cone's own breakup is not damped at all. This is tricky as you think your breakup is solved because measured frequency response is flat (when driving coil), while actually external perturbations are still going to trigger that ringing since the cone itself is not damped.
It would not be the first time I see drivers being stimulated by sources other than its own voice coil.
This could for example come from tweeter radiating energy close to mid, or mid's own distortion creating content landing on that breakup.
Can we buy the mid/tweeter driver separately? I d be interested to attempt a diy design.
Along with the “lossy interface between voice coil and midrange cone” (aka Neck Control) there is another trick used by KEF to suppress midrange driver break up and improve pistonic behavior: those ribs stamped on the cone itself improves structural strength.

"The stiffness is increased by the radial embossing in the cone profile, but nevertheless it is prone to high-Q breakup in the frequency range covered by the tweeter.” (that’s the reason neck control is also needed to suppress any possible out of band break up) All of this stuff is overly explained in the Blade Concept Whitepaper for whoever is interested.

One could say if we compare KEF Engineers work with the common audio company competitor and several DIY kits...this guys have made their homework and know their stuff.

By the way, the graph uploaded by @thewas_ corrsponds to The Reference line, this one is from 2018 R series UniQ:
E51C246F-B8DF-4444-8780-6FFBFFD83A85.jpeg

Just remember this is a 1300£/1600€/2000$ speaker (price could be much lower with discounts and Black Friday). I think what you get was simply inconceivable a few years ago at this pricepoint. Lets move on
 
Joined
Nov 4, 2019
Messages
25
Likes
14
Location
Belgium
One could say if we compare KEF Engineers work with the common audio company competitor and several DIY kits...this guys have made their homework and know their stuff.
No doubt about that. I never said R3 was a bad speaker. On the contrary, I almost bought them after doing my research, and that's saying something..

It measures great and I wanted to see if more detailed/specific measurements were needed to reveal reported issues. Turns out that the answer might have already been in there with the off axis responses, surprising since they looked near perfect in Amir's measurements.

We are discussing the science behind audio transducers on ... a forum about audio science. To me that's what makes ASR special place where we see great contributors. My excuses if you find this Out of Topic or annoying.

*prays the gods of science to remain in ignorance.
 

dwkdnvr

Active Member
Joined
Nov 2, 2018
Messages
242
Likes
314
I don't think anyone has claimed this (in the last comments).
It was said that "Some mention that they are not very subtle, and a bit harsh...". I then went into this, using the measurements as a basis.



Interpretation of measurements is anything but "sound-la-la-land", because this is exactly what measurements are made for. The interpretation I made is actually quite obvious if you look at the in-room responses of the R3 shown by @napilopez in post#866. There is in the mentioned range around 2-3kHz the in-room response a bit too prominent.
Must everyone see it that way? No, but my argumentation is based on measurements from different sources.

View attachment 84447


Take Two:

Okay,... anecdotal evidence is stupid, I got that.

;)
Those measurements are interesting. I'd still say that they are best approached not by cutting at ~2.8k, but by bringing up the 800-2.5k region. You might need a very slight cut at the peak, but bringing up the dip would appear to result in nice flat downward tilt.

I have a pair of R3s that I had used on my desktop where I quite liked them. When I tried them in the main space (a large open-concept area), the narrow(er) dispersion sounded a bit dull compared to what was there before (flat tapered baffle speakers). We're in the process of moving, and I have an opportunity to just go buy something like a pair of Revels, but am going to try the R3 + MiniDSP SHD + subs first. I'm intrigued as to what I find.
 

ctrl

Senior Member
Joined
Jan 24, 2020
Messages
495
Likes
1,670
Location
DE
Those measurements are interesting. I'd still say that they are best approached not by cutting at ~2.8k, but by bringing up the 800-2.5k region. You might need a very slight cut at the peak, but bringing up the dip would appear to result in nice flat downward tilt.
No objection on my part. The solution that sounds better is the better solution.

When using in-room measurement to check the result at the listening position, always keep in mind that the measuring microphone does not distinguish between vertical and horizontal radiation, but the human ear does to a considerable degree.
An even in-room FR does not automatically mean that the ear feels the same - and vice versa.
 

LarsS

Major Contributor
Forum Donor
Joined
Mar 23, 2016
Messages
1,049
Likes
962
Location
Sweden, Stockholm
Whenever it is reported that a loudspeaker is supposed to sound "a bit harsh", distortions are always considered to be evildoers.
Even if distortion measurements do not give any clues.

Warning: Anecdotal story of an old man.
Years ago I tested different tweeters and midrange drivers in a loudspeaker that sounded harsh above a certain sound pressure level. The harsh sound was present at similar axis frequency response independent of the driver and this although HD as well as MD (multi-tone distortion) was always very low.

The frequency range 2-5kHz is especially critical and there the R3 shows a small problem in horizontal and vertical radiation.

With non-coaxial loudspeakers, the crossover frequency to the tweeter is often in the 2-3kHz range. This means that "sound energy" is destroyed vertically in this range by cancellation, which can compensate a little bit for a horizontal widening of the radiation - this is not possible with coaxial loudspeakers.

The R3 now shows a widening in radiation in the range of 2-3kHz both horizontally and vertically, which may cause the sound to become somewhat harsh at higher sound pressure levels - if this widening is not completely compensated by the axis frequency response (the axis frequency response should show a slight dip in this range).

So I would rather see the radiation as a possible problem than the influence of distortions.

Except the R3 would sound harsh even outdoors or in very large rooms with no close boundary surfaces, which would indicate distortions as a cause.

View attachment 84432
Just out of curiosity, had KEF XQ40 a few years back. Sold after having had them less then a year for reasons similar what’s been discussed here. Harsh sounding when SPL is a bit high and bland.
Found this measurement of the XQ40’s in a German mag. Not skilled enough make comparison between this measurement and the R3. Appreciate any insight on any similarities or not.

1600968461652.jpeg


1600968065614.jpeg
 
Last edited:

ctrl

Senior Member
Joined
Jan 24, 2020
Messages
495
Likes
1,670
Location
DE
Just out of curiosity, had KEF XQ40 a few years back. Sold after having had them less then a year for reasons similar what’s been discussed here. Harsh sounding when SPL is a bit high and bland...
Not skilled enough make comparison between this measurement and the R3. Appreciate any insight on any similarities or not.
So I am by no means the "Doctor Dolittle" of speakers ;) Many statements can be justified with measurements, but sometimes you don't grasp the full complexity and it's no different than peddling fairy tales - keep that in mind.

Compared to many other loudspeakers, the radiation of the XQ40 is very even.
In this case I would pay attention to the range 2-4kHz and 4-6kHz.
2-4kHz: Are there two dips (which our brain hardly perceives) or a peak (in a wide dip)?
4-6kHz: This range provides a little more "sound energy" than the 2-4kHz range.

Now peddling fairy tales: I could imagine, for example, that in typical listening rooms distorted e-guitars and tightly struck cymbals might sound unpleasant at high sound pressure levels.

But, if you had said that the speaker sounded very pleasant and warm, I could have "justified" that with the measurements as well o_O

Predicting the sound on the basis of measurements is (for loudspeakers with good measurement results) extremely difficult (often 0.5dB in the high frequency range decide between "meh!" and "wow!"), justifying a listening impression with the measurements is easier ;)

Therefore it is not possible for me to tune a loudspeaker without listening - but the measurements help to find possible sources of error.
 

LarsS

Major Contributor
Forum Donor
Joined
Mar 23, 2016
Messages
1,049
Likes
962
Location
Sweden, Stockholm
So I am by no means the "Doctor Dolittle" of speakers ;) Many statements can be justified with measurements, but sometimes you don't grasp the full complexity and it's no different than peddling fairy tales - keep that in mind.

Compared to many other loudspeakers, the radiation of the XQ40 is very even.
In this case I would pay attention to the range 2-4kHz and 4-6kHz.
2-4kHz: Are there two dips (which our brain hardly perceives) or a peak (in a wide dip)?
4-6kHz: This range provides a little more "sound energy" than the 2-4kHz range.

Now peddling fairy tales: I could imagine, for example, that in typical listening rooms distorted e-guitars and tightly struck cymbals might sound unpleasant at high sound pressure levels.

But, if you had said that the speaker sounded very pleasant and warm, I could have "justified" that with the measurements as well o_O

Predicting the sound on the basis of measurements is (for loudspeakers with good measurement results) extremely difficult (often 0.5dB in the high frequency range decide between "meh!" and "wow!"), justifying a listening impression with the measurements is easier ;)

Therefore it is not possible for me to tune a loudspeaker without listening - but the measurements help to find possible sources of error.
Thanks for the insights!
Replaced the XQ40’s with these and quite happy with them.
1601013353054.jpeg
 
Joined
Mar 28, 2020
Messages
84
Likes
43
Do you guys think the CI160ER would pair well with the R3/5/7s in a 9.1.6 system? They seem to be dirt cheap right now and I don't know why. I don't know anything about in-wall speakers unfortunately.
 
Joined
Mar 28, 2020
Messages
84
Likes
43
Do you guys think the CI160ER would pair well with the R3/5/7s in a 9.1.6 system? As a follow-up I did track down these measurements. It isn't much but maybe someone smarter than me can comment on whether or not these are decent in-wall/in-ceiling speakers?

https://www.shop.us.kef.com/pub/media/wysiwyg/documents/ciseries/Ci160ER-EDIT-NO-CROPS.pdf
I was also able to find the measurements for the higher end THX certified models:

https://www.shop.us.kef.com/pub/media/wysiwyg/documents/ciseries/Ci160RR-THX-EDIT-30-01-20.pdf

Are these significantly better? I'm not familiar with the presentation format of directivity index.
 

BYRTT

Addicted to Fun and Learning
Forum Donor
Joined
Nov 2, 2018
Messages
796
Likes
1,975
Location
Denmark (Jutland)
I was also able to find the measurements for the higher end THX certified models:

https://www.shop.us.kef.com/pub/media/wysiwyg/documents/ciseries/Ci160RR-THX-EDIT-30-01-20.pdf

Are these significantly better? I'm not familiar with the presentation format of directivity index.
Much better in those huge dips for the previous model in directivity index curve for 5 and 10kHz areas is now much less, to understand what happens when directivity index on objective paper show ripple resonances is related to how smooth responses is as we go off axis relative to on axis, for example see below animation for R3 where we see a relative beatifull smooth directivity index curve and for the lower graph we see the nine axis's that when avaraged form listening window curve in upper graph, if directivity index curve is having trouble or much ripple then for that lower graph curves will begin look like a mess with spread around in the curves especially where those ripple resonances is presented into directivity index..
Snoochers_x1x1_500mS.gif
 
Joined
Mar 28, 2020
Messages
84
Likes
43
Ah so as usual a smooth line is better, eh? So the "R" THX model is significantly better than the "E" model. That doesn't surprise me because they're like 10x more expensive haha, but I was hoping they'd be more similar.

Why does the line only start going up at like 600hz in the chart I shared? The directivity index (red) has a lot more variation down all the way to 50hz in these R3 charts.
 

BYRTT

Addicted to Fun and Learning
Forum Donor
Joined
Nov 2, 2018
Messages
796
Likes
1,975
Location
Denmark (Jutland)
Ah so as usual a smooth line is better, eh? So the "R" THX model is significantly better than the "E" model. That doesn't surprise me because they're like 10x more expensive haha, but I was hoping they'd be more similar.....
Year as usual a smooth line is better and sorry it often means more expensive, we in theoretical and objective land so cant say what it means subjective when they get bandpass filtered in a 9.1.6 system, if one really can subjective notice going with the cheap or expensive one or some other solution.
.....Why does the line only start going up at like 600hz in the chart I shared? The directivity index (red) has a lot more variation down all the way to 50hz in these R3 charts.
Not that i know but will think the difference is because R3 is full space also called 4PI measured all 360º around in Amir's Klippel scanner and those KEF cuvers is half space also called 2PI measured 180º around because the wall mounting (baffle) blocks or hinders full space (4PI) acoustics to happen for the unit.
 

Similar threads

Top Bottom