• 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 R3 Speaker Review

Martigane

Member
Joined
Nov 4, 2019
Messages
45
Likes
64
Location
Belgium
we see extra energy above the bass from 4-10 kHz, which is very broad and the likely source of complaint.
Indeed, but that excess of energy is very small in level all things considered. Even it it was, it would ressemble the widely appreciated BBC dip which many correlate with nice low SPL listening balance, which makes sense, but I ve also seen several review stating that R3 needed to be pushed a bit to really sound good/balanced.
it will exacerbate any potential brightness in the upper mids and lower trebs.
true, but directivity and CSD are very good, so it makes me wonder if there are other aspects we are missing and should be measuring.
Could the inter modulation distortion due to the coaxial design be a source of issue? I am not sure the 32 multitone is the best test to highlight this. Maybe a 200Hz + 3kHz to focus on the worst case excursion / tweeter levels?
 
Last edited:

sfdoddsy

Active Member
Joined
Jun 6, 2019
Messages
293
Likes
438
The other hypothesis is that whilst measurements are objective, listeners are not.

Some simply may not like the sound of an accurate speaker.
 

tjkadar

Member
Joined
Mar 20, 2020
Messages
61
Likes
160
Location
Stateburg, South Carolina
Hi, in these graphs it would help to use a scale of 50-60db otherwise everything looks too flat

Kachda,

Here is another sweep in the preferred scale.

KEF R3 Sweep.jpg

I've already addressed the hump between 10 and 20 Hz with a different crossover slope but didn't save the measurement. Flattened it out nicely.

I built a pair of 12" sealed subs powered by a pair bridged Crown XLS 1502s to pair with the KEFs. I'm using a MiniDSP SHD for DSP duties now. Waiting on some GIK panels to help with a few room issues and a pair of March Audio P451 monoblocks.

I bought the KEF R3s based on Amirm's measurements and review. The KEFs measure well and, in my opinion, the listening experience backs that up.

I had the opportunity to audition the Magnepan 30.7 for a couple of hours on a wide variety of music. These were setup by the manufacturer and auditioned in a room specifically treated with the Maggies in mind. While I found the Maggies entertaining, they did not seem significantly better than my system. Especially on the lower registers. They did not have the bass I crave (but I am a bit of a basshead).
 

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,146
Likes
8,715
Location
NYC
Indeed, but that excess of energy is very small in level all things considered. Even it it was, it would ressemble the widely appreciated BBC dip which many correlate with nice low SPL listening balance, which makes sense, but I ve also seen several review stating that R3 needed to be pushed a bit to really sound good/balanced.

true, but directivity and CSD are very good, so it makes me wonder if there are other aspects we are missing and should be measuring.
Could the inter modulation due to the coaxial design be a source of issue? I am not sure the 32 multitone is the best test to highlight this. Maybe a 200Hz + 3kHw to focus on the worst case excursion / tweeter levels?

I highly doubt IMD has much to do with it in a three-way design.

What I'm saying is that the specific problems most people have described here are visible only from the frequency response. Also as far as I'm aware, there's nothing to suggest the BBC dip is in fact 'widely appreciated.'

Look at my post showing the problems with the 1kHz dip. I'll reshare my own measurement:

1600727558228.png


Here is the again the on-axis, my PIR calculation, as well as my actual in-room response at 2m:

R3 dip.png


and here is my PIR superimposed over another member's actual measured in-room response:
1600728365308.png


Depending on where you want to draw the target lines, the KEF R3 either has too little midrange energy or too much presence energy. I suspect that bass balance in your particular room could determine what you hear it as, since it seems we tend to anchor frequencies around 200-400hz.

I say this without meaning to sound curt, but I think you underestimate the differences small changes in the on-axis and directivity can make. For sure it would be fun to look at the rest of the data, but occam's razor points us to FR and directivity.

And I say all this while still believing the R3 is one of the very best passive speakers you can buy :).

The other hypothesis is that whilst measurements are objective, listeners are not.

Some simply may not like the sound of an accurate speaker.
That kind of goes agains the whole point of measurements and blind tests though. The whole point is of Harman's blind studies is to show that people overwhelmingly prefer speaker that tend towards neutrality in the direct and off axis sound.
 
Last edited:

sfdoddsy

Active Member
Joined
Jun 6, 2019
Messages
293
Likes
438
Yes, but obviously not everyone does.

I personally prefer less bass than the Harman curve, for example.
 

Martigane

Member
Joined
Nov 4, 2019
Messages
45
Likes
64
Location
Belgium
Depending on where you want to draw the target lines, the KEF R3 either has too little midrange energy or too much presence energy.
you are right, that dip is very wide!

I thought I saw a pretty flat response. maybe not so much :D
 
Last edited:

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,146
Likes
8,715
Location
NYC
you are right, that dip is very wide!

From stereophiule measurements I saw a pretty flat speaker. maybe not so much :D
As far as I know stereophile has not measured the R3! Which measurements are you talking about?
 

stevenswall

Major Contributor
Forum Donor
Joined
Dec 10, 2019
Messages
1,366
Likes
1,075
Location
Orem, UT
Allright. so now we got a loudspeaker that is measuring great - Kef R3.
Can ASR now explain why it gets so many bad reviews/feedback?
For me that is something ASR should dive into if objectivist are to claim measurements show all we need to know about a device's performance.
And I'm all for it!
Are we missing some measurements? Should we be looking at time domain distortion? CSD looks clean, only bass reflex in the bass looks slow, but that's just the bass and would not explain why there is so much negative feedback about the R3s.

What bad reviews are you seeing? What do they say about them?
 

Martigane

Member
Joined
Nov 4, 2019
Messages
45
Likes
64
Location
Belgium
What bad reviews are you seeing? What do they say about them?
I see lots of users selling them shortly after they bought them and being very disappointed. It might have to do with the reviews over-selling their bass, or with something else.
Some mention that they are not very subtle, and a bit harsh, and ... hold on... "lifeless". I can understrand that hard metallic cone has a sound signature of its own, but I do not see this reflected in the measurements (assuming cone breakup is already being taken care of by 20dB+ attenuation below main frequency response SPL)
It looks like a well engineered loudspeaker, but I can't shake the idea that we are missing something here.

Across the different loudspeaker candidates for my upcoming upgrade, this one popped up with this particular trend.
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,895
Likes
16,896
I can understrand that hard metallic cone has a sound signature of its own, but I do not see this reflected in the measurements (assuming cone breakup is already being taken care of by 20dB+ attenuation below main frequency response SPL)
KEF has done alot of research and tricks to suppress that breakup in the last few generations of the Uni-Q and also there are no complaints about it at their other model series. My guess is that most complaints about the R3 come about its tonal voicing choice (I also wasn't impressed too much when I listened to it), but that is nothing that an EQ cannot fix.
 

Hayabusa

Addicted to Fun and Learning
Joined
Oct 12, 2019
Messages
833
Likes
575
Location
Abu Dhabi
Allright. so now we got a loudspeaker that is measuring great - Kef R3.
Can ASR now explain why it gets so many bad reviews/feedback?

Measuring great with a speaker is still far from being transparent, so many aspects of these measurements show results in the audible range (FR fluctuations, distortion, directivity etc,)
So I am not surprised people have different opinions on a 'great measuring' speaker. The measurements support these differences.
 

Objectivist01

Senior Member
Joined
May 24, 2020
Messages
317
Likes
118
Allright. so now we got a loudspeaker that is measuring great - Kef R3.
Can ASR now explain why it gets so many bad reviews/feedback?
For me that is something ASR should dive into if objectivist are to claim measurements show all we need to know about a device's performance.
And I'm all for it!
Are we missing some measurements? Should we be looking at time domain distortion? CSD looks clean, only bass reflex in the bass looks slow, but that's just the bass and would not explain why there is so much negative feedback about the R3s.
most reviews are from idiots who had been living with colored sound all their life. Hope zhat explains. I have listened to this speaker and its really good. Better zhan most stuff at its price
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,454
Likes
15,806
Location
Oxfordshire
Allright. so now we got a loudspeaker that is measuring great - Kef R3.
Can ASR now explain why it gets so many bad reviews/feedback?
For me that is something ASR should dive into if objectivist are to claim measurements show all we need to know about a device's performance.
And I'm all for it!
Are we missing some measurements? Should we be looking at time domain distortion? CSD looks clean, only bass reflex in the bass looks slow, but that's just the bass and would not explain why there is so much negative feedback about the R3s.
IME most of the traditional hifi magazines, and hence a lot of web reviewers too, give primacy to price, so nothing as inexpensive as the R3 can get a top recommendation.
 

Frank Dernie

Master Contributor
Forum Donor
Joined
Mar 24, 2016
Messages
6,454
Likes
15,806
Location
Oxfordshire
I can understrand that hard metallic cone has a sound signature of its own
Metallic cones can be pistonic in their pass band, though eventual break up can be high amplitude narrow peaks but can be removed by the crossover.
The idea that a hard metalic cone could have a sound signature is as wrong as thinking silver cables should sound brighter than copper ones.
In fact most cone materials other than metal and ceramic will have (damped) cone breakup in their pass band so paper and plastic will have a characteristic sound.
 

sfdoddsy

Active Member
Joined
Jun 6, 2019
Messages
293
Likes
438
IME most of the traditional hifi magazines, and hence a lot of web reviewers too, give primacy to price, so nothing as inexpensive as the R3 can get a top recommendation.

Speaker of the Year at What HiFi, rated Outstanding at HiFi News.
 

tecnogadget

Addicted to Fun and Learning
Joined
May 21, 2018
Messages
558
Likes
1,012
Location
Madrid, Spain
What bad reviews are you seeing? What do they say about them?
I second this and have not seen that pile of " bad reviews" either.

I see lots of users selling them shortly after they bought them and being very disappointed. It might have to do with the reviews over-selling their bass, or with something else.
Still, have not seen lots of users selling their R3's, and I'm a member of the Facebook official Kef owners group, routinely search on local selling Apps and Hifishark.com. As Napilopez said, this can be explained by the usual counterculture attitude, speaking up against the few ones that got a high score and good measurements.

But mostly I think this mismatch in perception comes from 2 factors:
  1. People are VERY used to listening to colored sound signatures ("spicy sound"). Once they listen to something more neutral they feel is lifeless or dull...but I'm sorry, they are wrong. This kind of user usually has zero experience with calibration, house curves, room acoustics interaction, parametric Eq, etc. What happens is that they have some great expectations and think they will enter the Valhalla if they upgrade their low end-mid end speaker to this well-measured one. They start playing music and an audiophile vortex has not opened to their ears...they are listening to things much more correctly than before, the problem is they don't know that to listen for, they come from appreciating high distortion, uneven directivity, high discrepancies in frequency response and power response that "spice up" the music.
  2. Users that have zero experience with calibration, house curves, room acoustics interaction, parametric Eq, etc; enter ASR because they are music/sound "enthusiasts" and discover this speaker. They see this is objectively a great speaker and buy it with high expectations, take it out of the box, place it in their room, and think that the job is done. No wonder maybe something is wrong when you are having extreme comb filtering, SBIR issues, room modes excited, not enough separation between the speakers, etc. With this particular speaker, those are not real issues because of great directivity, but bass room modes are inherent to any speaker.
Those are the usual suspects you will find endlessly recommending a $6000 amp to "take the most" out of your LS50. Or asking why people use dual subwoofers. How much a $500 cable made their LS50 sing correctly, etc....You have seen them, I have seen them.

Some mention that they are not very subtle, and a bit harsh, and ... hold on... "lifeless". I can understrand that hard metallic cone has a sound signature of its own, but I do not see this reflected in the measurements (assuming cone breakup is already being taken care of by 20dB+ attenuation below main frequency response SPL)
It looks like a well engineered loudspeaker, but I can't shake the idea that we are missing something here.

Again, "hard metallic cone" is just a subjective perception people I did describe above came up. One of the measurable factors that describe this is when the dome breaks up near 20kHz or below so it becomes audible. As you have noticed there are usually Zero backings in those claims and there is nothing in Amir's or other site measurements that showings this. Kef Whitepaper kindly describes how the tweeter dome is constructed with special double-layer geometry and break up is well beyond 20khz.
1600768116785.png


That measured in-room response uploaded by Napilopez is from my living room, the most irregular room you could find, and speakers not properly placed yet with low separation between them. Still, I consider they measure excellent given the cited constraints. And it strikes me how I can walk all around the living room and even enter some sort of hallway and notice the special evenness in response. I walk all over the place and the tonal balance past 500Hz change is very low and subtle. This is something I have not experienced with normal 2-way or 3-way designs.

Another thing to consider with Concentric-Coaxial designs and good DI speakers is that what you see On-Axis on Amir's measurements won't directly translate to what you hear. PIR and Power Response match much better my In-Room measurements. So all that fuzz about elevated treble kind of fades away once you notice that past 2kHz the treble gently rolls off.
The famous 1kHz dip shows up on the measurement but very subtle. What makes it more visual is that below you have elevated bass (room interaction-build up) and above there is a small 2kHz peak. This places the 1khz dip at the bottom of an imaginary "V" letter. All of this can be and SHOULD BE corrected with PEQ.

Most of this talk is more technical or academic than anything...at the end of the day all I listen is good sound even with the given constraints and the only thing that sometimes could stand up are the room modes: the peaks and dips in the bass response below especially below 100hz where the fluctuations are at its maximum. And even then this isn't that much problematic since the hearing system is smart enough to let you filter this out, like some little voice in your head telling you "yeah I know those room modes are present, they are inherent to any speaker in a room, I'll forget you for being too lazy no not address it yet with propper PEQ, but I know you will do it sooner or later..."

Now we are coming back to point 1 & 2. These speakers have a lot of bass extension and clean output for their size, rivaling many floor standers. This is a double side sword. If you care about placement and PEQ for those bass peaks and dips you are on the clear. But if you are not going to take care of any of this, this speakers are many times more prone to excite the room modes than smaller speakers.

That's the dichotomy of Hi-Fi audio when sometimes you are fine with small or cheap speakers, start to upgrade to serious gear and disappoint because you are either not that engaged listening to a correct response, or you don't get the desired results because you don't spend the proper amount of time to propper set up the system of educating yourself about all the acoustic phenomenon involved in sound reproduction.

I would take any day 1 positive review/opinion from someone with credentials than 100 bad reviews from clueless users who saw a speaker measured great, bought it, and did not like it because didn't know what to look or care for. But that's just my opinion.
 

Attachments

  • 1600765423298.png
    1600765423298.png
    696.1 KB · Views: 133
Last edited:

Snoochers

Active Member
Joined
Mar 28, 2020
Messages
187
Likes
70
I'm interested in finding out the max SPL for the KEF R3 or R5 or R7 along the frequency response, similar to this Neuman graph. Is there such a thing?

neumann_kh310_max_spl_510_2.gif
 
Top Bottom