KEF R3 Speaker Review

[muad]

Oh yeah absolutely, same here. If it wasn't for this site, there's no way I could navigate all the noise from subjective reviews and manufacturer marketing hype. After finding this site and using these measurements to guide me, I really have not made a single bad audio purchase ever again.



Again, you are not the only one. There are many of us who found this. What I'd love to know is if this is due to the frequency response we see in the measurements, or if it's something else. To that end, we'd need someone with R3's and the ability to precisely apply DSP to fix the midrange recession, and see if that changes things dramatically.

I can say that if I artificially introduce a few db recession in the same frequencies of my favorite speakers, they also become dull and boring sounding, in similar ways to how most people describe the R3. So it's quite possible this could explain it.

The mid-range dip goes away if you push the flange and shadow flare in. I measured the in room resonse and it didn't have that 1khz dip that shows up in the majority of in room repsonses I've sent onlone. It still has a broad elevation in the highs that I perceive it as too much treble. I eq'd the speaker to beings the highs inline. Probably the most balanced thing I've ever heard. I feel it's a better speaker than the F206 and BMR that I previously owned. Really sounds incredible. I'm still going to sell them as I listen to too many garbage recordings and I own a selah speaker that's far less accurate, and kinda smoothes everything over. I don't feel my current speaker is in the the same league as the KEFs

 

[TimVG]

OK so i focused on mid-range:
1) Anne Sofie von Otter - Baby plays around
EQ off - start of the song details like lip smacks, teeth and tongue noises, "t" and "s" are overpowering
EQ on - i still hear all the details but her voice is fuller and more balanced, less bright

2) Nina Simone - I put a spell on you
EQ off - again saxophone and singers voice more in upper registers, more bright and also metallic
EQ on - no detail loss, balanced, think the tone is more accurate

3) Tom Waits - I hope that i dont't fall in love with you
EQ on - imaging is better, more space between guitars on the left/right and voice in the dead center, while without EQ voice is more mixed in with guitars and more in the back.

3 songs is not much but it's is a noticeable improvement.

It's a good start. If I read right you're waiting on a UMIK? Once you get that low range tuned-in things should only get better.

 

[Chromatischism]

The mid-range dip goes away if you push the flange and shadow flare in. I measured the in room resonse and it didn't have that 1khz dip that shows up in the majority of in room repsonses I've sent onlone. It still has a broad elevation in the highs that I perceive it as too much treble. I eq'd the speaer to beings the highs inline. Probably the most balanced thing I've ever heard. I feel it's a better speaker than the F206 and BMR that I previously owned. Really sounds incredible. I'm still going to sell them as I listen to too many garbage recordings and I own a selah speaker that's far less accurate, and kinda smoothes everything over. I don't feel my current speaker is in the the same league as the KEFs

If you want tonal accuracy but you also listen to "garbage" recordings, order up a pair of Buchardt S400's. Much easier to listen to.

 

[aarons915]

The mid-range dip goes away if you push the flange and shadow flare in. I measured the in room resonse and it didn't have that 1khz dip that shows up in the majority of in room repsonses I've sent onlone. It still has a broad elevation in the highs that I perceive it as too much treble. I eq'd the speaer to beings the highs inline. Probably the most balanced thing I've ever heard. I feel it's a better speaker than the F206 and BMR that I previously owned. Really sounds incredible. I'm still going to sell them as I listen to too many garbage recordings and I own a selah speaker that's far less accurate, and kinda smoothes everything over. I don't feel my current speaker is in the the same league as the KEFs

My original measurements showed a worse dip than what is there as well and after ensuring my shadow flare was seated, my new ones don't show it so yes it's very important for that to be fully seated but the measurements we see on this site don't appear to suffer from that issue and the mids are still slightly recessed, easily fixed by EQ. The previous R series was the same way so it seems intentional on KEF's part. Without EQ and the shadow flare seated they actually sound pretty good on their own now but they're still better with some EQ.

 

[Chagall]

Agreed.

The room was not optimal by any stretch and the sales personnel weren't going to let me run the YPAO room EQ then and there. Luckily though it was around 1:30 in the afternoon on a Tuesday or whatever, the place was deserted and once the manager realized I knew my way around a Yammy AVR, he pretty well left me to my own devices to try different DSP settings/modes & whatnot.

I decided not to purchase them based on what I was hearing in Pure Direct mode. I wanted them to be great, I really did. The sales price was very affordable ($1,400 + shipping) and the seller had just recently purchased them 5 months ago (he sent me a pic of the sales receipt) and claimed he wanted to upgrade to a KEF tower.

Who knows maybe I could have manipulated the YPAO to my tastes.

But connecting better speaker to cheapo avr's that surely can't sound good.
The logic of it completely escapes me.
Random avr - 5.1 - 600 euro vs integrated stereo 600 euro.
How much of that 600 euro is for dolby/thx/dts certifications and stickers, lcd screen, radio tuner, hdmi and other connections, YPAO mics and how much for that five amp?

 

[tecnogadget]

The mid-range dip goes away if you push the flange and shadow flare in. I measured the in room resonse and it didn't have that 1khz dip that shows up in the majority of in room repsonses I've sent onlone.

Great ! Would you be so kind to post your in room measurements when you got some time ? Thanks

I’ve found its easy to pull out and reposition both Shadow Flare and woofer trim ring with an iPhone opening tool like this:

DISCLAIMER ! Never use anything hard&sharp like metal tweezers or screwdriver. Always make sure to avoid the driver unit and be gentle.Use only soft tools made of plastic plus common sense.

But connecting better speaker to cheapo avr's that surely can't sound good.
The logic of it completely escapes me.
Random avr - 5.1 - 600 euro vs integrated stereo 600 euro.
How much of that 600 euro is for dolby/thx/dts certifications and stickers, lcd screen, radio tuner, hdmi and other connections, YPAO mics and how much for that five amp?

Not really. Unless the subject AVR is really broken (headless panther) you can get away with murder with midrange or cheapo old AVR’s, it won’t sound particularly bad.
Here we care about SINAD and all that stuff but in reality it’s just not night and day. The most obvious limitation with cheap (new and full price) $500 AVR will be weak amplification, specially with all channels driven. But hell I can understand somebody connecting their R3’s to that if it’s transitional, it will still sound great, even though the final bar is much higher.

 

[Kachda]

What are you guys using to eq/dsp ? I also own the r3 but dont have any eq tools at the moment. I drive them with a outlaw rr2160 and xover at 80.

 

[aarons915]

I've put together a filterset for you with respect to all of Amir's data.

View attachment 72577

It flattens the direct sound, taking in account the side-wall reflection and overall in-room curve

View attachment 72578

You can tell aren't just trying to make a nice flat line with your EQ, I'm sure you've found out from your F206 that you have to compromise between the different curves to get the best sound quality. I have the added bonus of using in room measurements and listening tests to compare but the filters I think I'm finally settled on are similar to yours. A couple of areas I ignored are the 370Hz peak and the roll off in the highs, my in room measurements didn't show them to be a problem and the highs actually measured almost identically to my LS50 after 7k, which don't show to be rolled off here or in the Soundstage measurements, so I left them alone. Also the area right around 5k levels out off-axis so it's not really necessary to reduce that region either, I confirmed that with listening tests.

We have similar philosophies though, I try to use minimal EQ to achieve a more neutral LW and ER response, it doesn't have to be dead neutral. The R3 starts out in a +/- 1 db window already, my filters reduce that to +/- .5db with a smooth ER curve but when comparing A/B it's not that big of a difference but the EQ version is definitely better and more natural sounding.

Anyway here are my filters and before and after graphs, I smooth them to 1/6 because I'm only worried about overall tonal balance:

Freq Gain Q
650 -.8 .6
1250 -.5 20
1640 1.7 .7
2700 -1 4
5250 -.3 1

 

[Chagall]

What are you guys using to eq/dsp ? I also own the r3 but dont have any eq tools at the moment. I drive them with a outlaw rr2160 and xover at 80.

You can use APO equalizer for windows with Peace plugin.
https://sourceforge.net/projects/peace-equalizer-apo-extension/

 

[Jas0_0]

Following the impact of low temperature noted in the Neumann KH80 revision posted today, I wondered whether temperature might also have contributed to the low frequency droop of the KEF R3, the review of which was posted in early March. I live in the UK where early March can be pretty freezing, not sure what it’s like where Amir is. Is temperature a potential factor?

 

[tecnogadget]

Following the impact of low temperature noted in the Neumann KH80 revision posted today, I wondered whether temperature might also have contributed to the low frequency droop of the KEF R3, the review of which was posted in early March. I live in the UK where early March can be pretty freezing, not sure what it’s like where Amir is. Is temperature a potential factor?

I had the same exact thought

But then noticed this also shows in every other KEF measurements (LS50, Q100, Q350) so I would say its more of an intended signature than the effects of temperature, but still have some doubts...

 

[Jas0_0]

I had the same exact thought
View attachment 72813
But then noticed this also shows in every other KEF measurements (LS50, Q100, Q350) so I would say its more of an intended signature than the effects of temperature, but still have some doubts...

Ah yes, good spot - it does seem to be a consistent characteristic and the Q350 was reviewed in May, so probably independent of temperature.

 

[tjkadar]

I'm using a MiniDSP with Dirac.

Here are my in room measurements. Green is a sweep with the microphone at the LP and yellow is using the moving microphone method. I am using two 10" subs to fill in the bottom end.

 

[tecnogadget]

@tjkadar That’s marvelous ! It surely has to sound really nice !

 

[tjkadar]

I think they reproduce music fairly accurately. Just a Little Lovin' by Shelby Lynne is amazing. Bird on a Wire by Jennifer Warnes, Noah by the Fairfield Four, Creole Moon by Dr. John, and more all sound wonderful. Plus, there is enough oomph to throw down something like Metallica's live version of Whiskey in the Jar'O at concert-loud levels! It is a good setup and extremely enjoyable.

 

[Kachda]

I'm using a MiniDSP with Dirac.

Here are my in room measurements. Green is a sweep with the microphone at the LP and yellow is using the moving microphone method. I am using two 10" subs to fill in the bottom end. View attachment 72838

Hi, in these graphs it would help to use a scale of 50-60db otherwise everything looks too flat

 

[Martigane]

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.

 

[HooStat]

Can ASR now explain why it gets so many bad reviews/feedback?

I think the first place to start is to understand whether this statement is even true. I think many people love them and I think they have pretty good reviews.

 

[Chromatischism]

I've seen several people subjectively say it's analytical and only good with high quality recordings. That would point me towards certain portions of the treble spectrum. In this case, we see extra energy above the bass from 4-10 kHz, which is very broad and the likely source of complaint.

 

[napilopez]

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.

I think the first place to start is to understand whether this statement is even true. I think many people love them and I think they have pretty good reviews.

The KEF R3 is one of the most popular speakers on the market and is extremely well-reviewed on the whole. I think ASR like many forums tends to have a bit of a counterculture attitude, in the sense that people seem to speak up more when it's against popular opinion so that I'd gather that's at least a bit of what's happening here.

Regardless, I think the R3 is very good. As I've said many times in this thread (I don't blame you if you haven't read the full thing), the majority of measured units show a noticeable on-axis dip at 1K, and arguably bit of an elevated response in the presence region. Amir's unit doesn't show the broad 1K dip, but almost every other measurement does. When it exists, it will exacerbate any potential brightness in the upper mids and lower trebs.

Those two things are what people have recurringly picked up on. I found the R3 slightly recessed in the mids, but others find it a bit bright (brightness did not come to mind for me, but it's one and the same issue). The reasons why are evident from the spin.

So regardless of how 'good' the speaker is, the measurements tell us what to expect when people pick up on a flaw with the speaker.