KEF R3 vs R5

[Kachda]

Are you saying having 2 5.25 inch woofers is the same as a 7 inch woofer? Not to be ignorant, but is there a specific ratio or formula you use to make the comparison of multiple small woofers to one larger?

Surface area formula I would assume.
2x5.25 area = 1x7.5. But these are approximations since you need to take into account sorround area

 

[BrokenEnglishGuy]

Are you saying having 2 5.25 inch woofers is the same as a 7 inch woofer? Not to be ignorant, but is there a specific ratio or formula you use to make the comparison of multiple small woofers to one larger?

this post is very interesting, i didn't do the correct maths, i did by memory but it's something like that i said

see this post, here are the maths
https://www.avforums.com/threads/re...-many-of-these-to-equal-one-of-those.1899701/

 

[Matias]

And table here:
https://www.engineeringtoolbox.com/circles-circumferences-areas-d_1346.html

 

[amicusterrae]

This is an interesting choice with the current sale--one I've thought about, too. I wish KEF disclosed the full spin data, as we only have DI for Amir's measurements of the R3.

The R5 doesn't seem to follow the typical bookshelf vs. tower tradeoffs. It isn't more efficient than the R3, and it doesn't seem to have more bass extension (to get some of those gains, you need to go to the R7). You're likely going to want dedicated subs with either the R3 or the R5. Also, unlike a lot of bookshelf speakers, the R3 is a three-way, so gaining a dedicated mid-range is not a reason to "step up" to the R5 (compare the Revel M106 vs. F206).

My guess is the R5 was intended to be the tower version of the R3. Some folks just don't prefer bookshelf speakers. I love the look of the R5 and would much rather have it in my living room than the R3 on stands.

Something else to consider. The tweeter on the R5 is not at the top of the speaker. In my room, it would placed well below seated ear level, and would need some base platform to lift it. With the R3, you need stands anyway for the best sound and probably can find the specific height for your room. To me, that made it easy to pass on the current R5 sale.

Finally, hear's another way to look at it. Based on available data, the R3 is one of the best performing speakers in the world (with a sub). KEF's data indicates the R5 is in the same class. There are some tradeoffs with KEF's unique Uni-Q array versus other designs, but between the R3 and R5, you probably can't go wrong!

 

[Matias]

Also consider that the R5 has woofers both above and below the coax, so imaging is likely to be better focused in middle height, compared to woofer below coax of the R3.

 

[BrokenEnglishGuy]

Also if the R5 can go low as R3 but with smaller woofers, it will be faster

 

[Pio2001]

I also pasted the spin data from the KEF white paper for both speakers so it is easier to see.View attachment 108192

The R3 data is clearly superior to the R5 because of the 3 dB drop at 3 kHz in the "semi-sound power" curve of the R5.
This drop should be audible in a standard room.

Also consider that the R5 has woofers both above and below the coax, so imaging is likely to be better focused in middle height, compared to woofer below coax of the R3.

The vertical directivity plot of the R3 is near-perfect.

Also if the R5 can go low as R3 but with smaller woofers, it will be faster

Is there any scientific evidence ? It sounds like a misconception to me.

 

[warthor]

The R3 data is clearly superior to the R5 because of the 3 dB drop at 3 kHz in the "semi-sound power" curve of the R5.
This drop should be audible in a standard room.

The vertical directivity plot of the R3 is near-perfect.

Thanks a lot for pointing out a difference in the curve. I was worried the R5 would not be quite as good as the R3, and that very well may be the case.

 

[Ron Texas]

With the R5 you don't need stands.

 

[thewas]

The R3 data is clearly superior to the R5 because of the 3 dB drop at 3 kHz in the "semi-sound power" curve of the R5.
This drop should be audible in a standard room.

I don't really see a real problem there, just higher directivity for the R5 (but also smooth) that can have even advantages at lively rooms and/or higher listening distances.

 

[Pio2001]

I don't really see a real problem there, just higher directivity for the R5 (but also smooth) that can have even advantages at lively rooms and/or higher listening distances.

I am not talking about the general slope, but about the small deviation between 3 and 4 kHz.
In the two graphs posted above, if we draw a straight line across the purple curve (power response across the front hemisphere) from 1000 Hz to 8000 Hz, which is the range where the speaker sound power is dominant, we can see that the R5 has a weak output at 3500 Hz, while the R3 is more neutral.

This might lead to a noticeable difference in the predicted (and actual) in-room response of the speaker, although these curves are only front hemisphere sound power. It would be interesting to see the full sound power measurement.

 

[Pio2001]

Here is the deviation that I am talking about.

It is small, but we don't know how it would look in the complete sound power measurement of the speaker.

 

[andreasmaaan]

Surface area formula I would assume.
2x5.25 area = 1x7.5. But these are approximations since you need to take into account sorround area

You also need to keep in mind that what is normally marketed as a "5.25-inch" woofer typically has the surface area of a circle of diameter 4" (approx. 12.6 sq inches), what is marketed as a "6-inch" woofer typically has the surface area of a circle of diameter 5" (approx. 19.6 sq inches), and so on...

 

[Koeitje]

The biggest difference is going to be distortion in the low end when playing louder.

 

[thewas]

I am not talking about the general slope, but about the small deviation between 3 and 4 kHz.

I see what you mean, please mind that this kink (due to change in slope from decreasing to constant) is often observed in loudspeakers with kind of constant directivity tweeters/horns, like for example on most highly regarded Harman designs, exemplarily

(at around 1,5 kHz)

(before 1kHz)

and from users experiences not really negatively noticeable.

 

[Zvu]

Here is the deviation that I am talking about.

It is small, but we don't know how it would look in the complete sound power measurement of the speaker.View attachment 108714

More detailed measurements Amir made show that you have to do a bit of equalization to get the best sound out of R3. Since you already need to equalize both of them, R5 is better option. 1dB-2dB at 3.5kHz solves it. If left unequalized, both speakers will have some flaws. R3 between 1-2kHz, R5 between 3.2-4.2kHz.

That being said, we haven't seen more detailed measurements of R5, so it could be better or worse - we can't know for sure.

 

[Alexanderc]

This might lead to a noticeable difference in the predicted (and actual) in-room response of the speaker, although these curves are only front hemisphere sound power. It would be interesting to see the full sound power measurement.

Based on your observation, I wonder how the rear ports would play into the total.

 

[warthor]

Here is the deviation that I am talking about.

It is small, but we don't know how it would look in the complete sound power measurement of the speaker.View attachment 108714

Thanks for pointing out these differences and characteristics on the spin graphs. Being new to these graphs I don't have the eye to see what you are seeing. But I can learn as you point them out (a little bit anyway).

I suspected we needed some measurements to be sure of the quality of the R5 (which is true in all areas of science), but it seems reasonable based your comments that we can expect quite a bit of similarities between the R3 and R5 (more similar than different).

I saw a steep drop at 7500 Hz (pictured) in the R5, that doesn't occur in the R3. Is that meaningful? It seems like a steep drop.

 

[aarons915]

Thanks for pointing out these differences and characteristics on the spin graphs. Being new to these graphs I don't have the eye to see what you are seeing. But I can learn as you point them out (a little bit anyway).

I suspected we needed some measurements to be sure of the quality of the R5 (which is true in all areas of science), but it seems reasonable based your comments that we can expect quite a bit of similarities between the R3 and R5 (more similar than different).

I saw a steep drop at 7500 Hz (pictured) in the R5, that doesn't occur in the R3. Is that meaningful? It seems like a steep drop. View attachment 108762

The 7500Hz notch happens only directly on axis so I wouldn't be concerned with that. As someone who has had the R3 for about 6 months now, the biggest problem with them in my opinion is the region in the 2-4k range, most of the R series has it in some way and it is only evident in the early reflections, meaning it isn't a problem in the listening window. The R3 and R5 would be fairly similar in most ways and I think it would come down to whether you prefer towers or bookshelf speakers but the R3 is slightly smoother according to KEF's measurements. The R7 is actually the smoothest to me in that region but I'm not sure how much of an audible difference that would make.

 

[Pio2001]

The 7500Hz notch happens only directly on axis so I wouldn't be concerned with that.

Agreed. This is a notch, not a peak. It is very narrow. And it doesn't exist unless you point the tweeter right towards your ear.

More detailed measurements Amir made show that you have to do a bit of equalization to get the best sound out of R3.

That's right. Amir's measurements for the R3 are not as forgiving as the one shown here, and a bit of eq would be perfect to get a flat frequency response.
The smooth sound power curve says "eq will work like a charm".

Since you already need to equalize both of them, R5 is better option. 1dB-2dB at 3.5kHz solves it.

As far as sound power is concerned, if there is any issue with the R5 (we can't tell because it already looks very good on both speakers), equalization won't solve it because equalization can't correct separately sound power and on-axis response.