KEF R11 Meta Tower Speaker Review

[HipHopPotatoMouse]

Thanks for all the responses, y'all!

@ban25 and @HeadDoc12 -- I agree with you and am leaning towards R5 Meta at this point. I wasn't noticing too much difference in the SPL charts of R5 vs. R11 anyways.

@Laserjock -- Oh man, you're now sending me into another rabbit hole Q11 Metas are also shorter, so they won't be obstructed by my weird furniture. I'll look a bit more into its comparison vs. R5 Metas, but Q11 may not be able to compete with the great deal I found for the R5

@Jumu -- My comment about lacking bass was actually wrt. music. With HT, sub is doing a wonderful job. But when it came to music, there was some gap. For example, I love Tool, and tested the R3s with Lateralus and in e.g. Disposition, or with E.S.T's From Gagarin's point of view, where the song is carried by bass guitar/ double bass, a lot of the body of the low bass notes was subdued.

I took a single point measurement w/ and w/o the sub for R3, and compared to an old measurement for my LS450s w/ sub (attached). What I find confusing with these measurements, is that, there is no clear deficiency above 40Hz in the SPL, but the feel is very different. I used graphic EQ to provide some boost between 50-100Hz (somewhat similar to what @amirm did) which recovered a lot of what I felt was missing.

I need to check, but I use the Dayton processor/room correction for the subwoofer, and I may have added a crossover at 80Hz for the subwoofer there. I may need to move that to 90 or 100 Hz to have a more sufficient compensation by the subwoofer.

 

[Laserjock]

@HipHopPotatoMouse measurements coming soon..

 

[exm]

This. I have the Reference 3 and even they are surprisingly lacking in bass probably because they only come with two 6,5" woofers. I think the limit is volume since the KEF are usually really sleek. But as long as you let a sub do a job, it's fine. If you want full range only from speakers you'd need to have huge towers.

I had the R11s and moved to the Reference 5. The Reference 5s are as close as I can imagine to the old Reference Fours I had in the past. Not sure if the Reference 3s have so much 'less' bass but the Reference Series 6.5" LF drivers + cabinet + crossovers are a league above the R series.

To OP: how about the R7 Meta?

 

[Jumu]

@Jumu -- My comment about lacking bass was actually wrt. music. With HT, sub is doing a wonderful job. But when it came to music, there was some gap. For example, I love Tool, and tested the R3s with Lateralus and in e.g. Disposition, or with E.S.T's From Gagarin's point of view, where the song is carried by bass guitar/ double bass, a lot of the body of the low bass notes was subdued.

I see that you already know everything you need. You don’t need information, you need advices, choices
My advice is pair R3 with a very low THD sealed sub, something like Perlisten D212(215) or D15 or R212, then you will have better results than towers.

 

[HipHopPotatoMouse]

I ended up getting the R11. It is integrating better with my sub, has provided the LF extension I felt I was missing in R3 metas, and I am equally enjoying the improved clarity compared to my old speakers. Thanks for all the discussion!

 

[Juhazi]

R11 would benefit from being close to front wall an no toe-in. On-axis dip ar 1200Hz is compensated by vertical dispersion in farfield room response.
A seriously good speaker and yes, as Center you can put it sideways below the screen/TV. It's important to use a center with similar xo topology as L/R (phase!)

Old version without Meta is ok too https://www.soundstagenetwork.com/i...&catid=77:loudspeaker-measurements&Itemid=153

 

[tetsu96]

price difference between towers and bookshelves for this nuances is too much.

Maybe, but it's nice to have towers that dig deep even if they're playing with subs. Clean bass from towers AND clean bass from the sub(s), AND in alignment? Absolutely fantastic.

The Reference 3 has hell lot of bass until 40hz . I demoed them against, elac vela 409, canton vento 90, canton reference 5k, focal quadral aurum vulkan 9, epos es28n some weeks back. it’s in the same ball park but like all speakers if you lack bass at your listening spot, try this: reverse the phase of one speaker to see if bass comes up.(you will loose the stereo image, but this is just for test) Then the room null is the problem. I have r11, before adding the corner trap I had zero bass from them. Literally nothing, I can see the woofers flying. I can feel my couch shaking, but I can’t hear any bass!!! One corner trap on one corner did the job for me. Somehow that corner was messing with the bass in my room due to its layout and furniture placement. This can happen to any speaker.

I get to 35Hz on my R11 Metas in room before the bass drops off, very little shortage of low frequency in those guys (and I run them LFE+Mains meaning they run full range and they're coupled with the sub).

Interesting how the room modes go - viewing audyssey captures, I can really appreciate how a mic 1ft to the left of the MLP shows a 10db drop at 45Hz, and it's +5db going 1ft to the right instead. This is only my front right speaker (left is more well behaved), and it doesn’t come off as obnoxious when listening. The average FR is mostly good.

Maybe I’ll do more than carpet / couch / blankets, but it’s a family room, so I probably won’t.

I ended up getting the R11. It is integrating better with my sub, has provided the LF extension I felt I was missing in R3 metas, and I am equally enjoying the improved clarity compared to my old speakers. Thanks for all the discussion!

I think that's a solid choice, and you'll love it even more as you dial in positioning / room correction / tuning.

 

[capslock]

Excuse me if this has been addressed before. How does this graph square with the a 330 Hz (some older data sheets say 400 Hz crossover frequency as specified by KEF?

This looks more like a 190 Hz -6 dB point = XO frequency to me.

Edit: I wondered if with the measurement below being nearfield the far field would look different due to baffle step. However, with its 20 cm baffle width, the drop from baffle step would start at a much higher frequency than either 190 or 330 Hz. I would also assume this was measured with the crossover in place, but it looks like the regular closed box 12 dB/oct drop off of a MF driver in a sealed compartment. Very confusing!

Bonus question: why is the XO frequency all over the place with R3, R5, R7 and R11, with no correlation to LF size or overall speaker size?

 

[capslock]

Anyone understand why the mid/high crossover point should be different in the different models?

R3: 420 Hz, 2.3 kHz
R5: 400 Hz, 2.7 kHz
R7: 400 Hz, 2.4 kHz
R11: 330 Hz, 2.5 kHz

So my bonus question has been asked in the general R Meta series thread but was never answered to any satisfaction: https://www.audiosciencereview.com/.../kef-r-meta-series-release.41420/post-1481928


Apparently, the R series from 2018 (without Meta) had the same crossover points throughout.

It has been stated (I didn't find the post at this point) that the Uni-Q is the same across the series (with the exception of the R8 Meta, of course, that lacks a woofer).

Let me explain where I struggle to understand.

The R3 and R11 have 6.5 in woofers (LF units in KEF speak) and a wider baffle and hence the baffle step is lower in frequency. A crossover should be easier to design if the XO point is chosen lower. Yet the R3 has the highest LF/MF point of the lot and the R11 the lowest. The R5 and R7 with their smaller woofers and narrower baffle are just in between the extremes.

What is more, the R11 with its four large woofers can generate the highest LF output. If the Uni-Q is indeed the same across the series, one would expect that a higher XO point and / or a steeper high pass would make sense. But, again expectedly, the R11 hat the lowest XO point and what looks like a leisurely 12 dB / Oct high pass with a - 3 dB point of about 220 Hz (see post before).

 

[Soniclife]

So my bonus question has been asked in the general R Meta series thread but was never answered to any satisfaction: https://www.audiosciencereview.com/.../kef-r-meta-series-release.41420/post-1481928


Apparently, the R series from 2018 (without Meta) had the same crossover points throughout.

It has been stated (I didn't find the post at this point) that the Uni-Q is the same across the series (with the exception of the R8 Meta, of course, that lacks a woofer).

Let me explain where I struggle to understand.

The R3 and R11 have 6.5 in woofers (LF units in KEF speak) and a wider baffle and hence the baffle step is lower in frequency. A crossover should be easier to design if the XO point is chosen lower. Yet the R3 has the highest LF/MF point of the lot and the R11 the lowest. The R5 and R7 with their smaller woofers and narrower baffle are just in between the extremes.

What is more, the R11 with its four large woofers can generate the highest LF output. If the Uni-Q is indeed the same across the series, one would expect that a higher XO point and / or a steeper high pass would make sense. But, again expectedly, the R11 hat the lowest XO point and what looks like a leisurely 12 dB / Oct high pass with a - 3 dB point of about 220 Hz (see post before).

Avoiding comb filtering issues from multiple drivers playing the same frequency that are spread out is my guess. If they added a 0.5 way to the crossover they would have more flexibility.

 

[capslock]

We get virtually the same picture as Amir got for the R11 Meta for the R3 Meta as measured by Nuyes: gentle 6 dB/oct low pass rolloff for the LF above 120 Hz and 12 dB MF high pass rolloff for the MF. I'd place the - 6 dB point at about 170 Hz. With this, it makes even less sense for me why they would call it 420 Hz in the R3 and 330 Hz in the R11. I will acknowledge that for the R11, likely only one of the four woofers was measured.

KEF R3 meta Measurements

Hi, I measured the KEF R3 meta loudspeakers this time. These speakers have just been released, but I was able to get a pre-order sample (a pair) from an audiophile in Korea. Impedance Frequency response I measure HF responses using a 5ms window gating. Therefore, the...

www.audiosciencereview.com

The R6 Meta measured by Nuyes https://www.audiosciencereview.com/...ds/kef-r6-meta-measurements-and-review.46470/
has a similar overlap, but adds a rising response on the Uni-Q. This may be owed to the fact the the R6 is a center that has a very horizontally very wide baffle. Ignoring this gentle rise, I'd place the -3 dB point for the MF at 150 Hz. The data sheet states 500 Hz. So we get a lower apparent XO point from the measurement compared to R3 and R11, but a higher stated point.

 

[capslock]

Avoiding comb filtering issues from multiple drivers playing the same frequency that are spread out is my guess. If they added a 0.5 way to the crossover they would have more flexibility.

Possible, but then they would apply additional low pass filtering only to the outer LF units. I can see nothing in their spec sheets or whitepaper that would hint at a 3.5 way.

 

[Soniclife]

Possible, but then they would apply additional low pass filtering only to the outer LF units. I can see nothing in their spec sheets or whitepaper that would hint at a 3.5 way.

I'm saying because they haven't used a 3.5 way crossover they have to lower the crossover frequency.

 

[capslock]

I'm saying because they haven't used a 3.5 way crossover they have to lower the crossover frequency.

You have a point. But then, why wouldn't that apply to the R7, too? And the best way really to avoid comb filtering is not to have a 6 dB low pass on the LF units in the first place.

 

[Soniclife]

You have a point. But then, why wouldn't that apply to the R7, too? And the best way really to avoid comb filtering is not to have a 6 dB low pass on the LF units in the first place.

The R7 is lower than the R3.

You would need to do the maths with the physical distances to see if the theory works.

 

[AOR]

Excuse me if this has been addressed before. How does this graph square with the a 330 Hz (some older data sheets say 400 Hz crossover frequency as specified by KEF?

This looks more like a 190 Hz -6 dB point = XO frequency to me.

Edit: I wondered if with the measurement below being nearfield the far field would look different due to baffle step. However, with its 20 cm baffle width, the drop from baffle step would start at a much higher frequency than either 190 or 330 Hz. I would also assume this was measured with the crossover in place, but it looks like the regular closed box 12 dB/oct drop off of a MF driver in a sealed compartment. Very confusing!

Bonus question: why is the XO frequency all over the place with R3, R5, R7 and R11, with no correlation to LF size or overall speaker size?
View attachment 419958

Hi @capslock,

Nearfield measurements are useful for isolating the effects of each individual driver, but they aren't quite correct in the crossover region. This is because many drive units will be sounding simultaneously, and the measurement will only be in the nearfield for one of those drive units. Our crossover specifications are based on sectional, anechoic farfield data. For the R11 Meta, this can be seen below.

This shows the true response of the drive units, including positional and radiative information (drive location and baffle step etc.), that the nearfields don't capture.

Nearfield measurement has become standard practice in measurement-based reviews, as it does show the behaviour of an individual driver without messing around with the crossover.

On the other point about the difference in crossover points, I can't comment from my own experience, as I didn't design this range. However, a lower LF/MF crossover point for the larger speaker makes sense to me. This would be due to the large LF "source size" - the drive units are more spread out. This gives a narrower radiation than for a closer spacing of drive units. Therefore to give a smooth transition to the midrange, the LFs must be crossed over lower than if the drivers were tighter positioned. This is why blade is so good, the drive units are close together, so we can cross substantially higher, while still getting the correct directivity profile.

 

[capslock]

Hi @capslock,

Nearfield measurements are useful for isolating the effects of each individual driver, but they aren't quite correct in the crossover region. This is because many drive units will be sounding simultaneously, and the measurement will only be in the nearfield for one of those drive units. Our crossover specifications are based on sectional, anechoic farfield data. For the R11 Meta, this can be seen below.

View attachment 420062
This shows the true response of the drive units, including positional and radiative information (drive location and baffle step etc.), that the nearfields don't capture.

Nearfield measurement has become standard practice in measurement-based reviews, as it does show the behaviour of an individual driver without messing around with the crossover.

On the other point about the difference in crossover points, I can't comment from my own experience, as I didn't design this range. However, a lower LF/MF crossover point for the larger speaker makes sense to me. This would be due to the large LF "source size" - the drive units are more spread out. This gives a narrower radiation than for a closer spacing of drive units. Therefore to give a smooth transition to the midrange, the LFs must be crossed over lower than if the drivers were tighter positioned. This is why blade is so good, the drive units are close together, so we can cross substantially higher, while still getting the correct directivity profile.

Thanks! We were lacking the relative contribution of the drivers in the far field from the Klippel measurements. Would you happen to be able to post a similar graph for the R3 Meta and one of the ones with smaller LF units (R5/R7)?

 

[AOR]

Thanks! We were lacking the relative contribution of the drivers in the far field from the Klippel measurements. Would you happen to be able to post a similar graph for the R3 Meta and one of the ones with smaller LF units (R5/R7)?

Sure.
R7 Meta

R5 Meta

R3 Meta

As you can see, the LFMF crossover point is higher up on the smaller speakers. However, the drive unit size is a bit different to how you describe. The R5 (and R2) have 5.25" LF drivers, whereas all the other speakers have 6.5" LFs.

 

[capslock]

Thanks. My bad about the driver size of the R7 Meta. So from the graphs, R11M at 330 Hz and R3M at 420 Hz as from their data sheets are spot on. I would interpret the others as 350 Hz for the R7M and 380 Hz for the R5M. Their datasheets both say 400 Hz, hence my confusion.

As you said, a consistent downwards move of the XO with increasing speaker size makes sense and is reflected in the graphs.

 

[mxmzd]

Hope you guys have some guidance on this topic.

I am planning a Home Cinema with R7 Meta, R6 Meta and just couldn’t help myself and got a pair of R5 (non-meta) as rears in perfect condition for 1k€. Seller told me he couldn’t tell the difference to the METAs in a blind test. My thinking was that saving >1000€ compared to the R3 Metas plus stands is worth it.

Well, now I found a pair of R11s (non-META) for 2600€. But that would be too wild, right?Room has 25sqm. Talk me out of this

R7 Metas will be 4000€ the pair (new).