KEF Reference3 vs b&w 804 D3

[KenM10759]

...and music. The SDA soundstage is insane, sounds are literally coming from your direct left and right. A lot of fun with some content. As for the low end, talk about full range...jeez- doors, rafters and pant legs rattle.



LOL no doubt, I hear you. At 50" tall and 142 lbs each they're not easy hide. I was lucky enough to be able to "temporarily" set these up in the finished basement.

My preference is speakers which sit in a room and you can't really detect that sound is coming from from them. Rather, they project a virtual soundstage with breadth and depth apart from speaker cabinets. So far, the KEF Blade 2 I've heard multiple times are the very best at doing it that I've experienced.

 

[AcuDefTechGuy]

Regardless, the the 805D are undeniably in an entirely different league over the CM5. I currently own both models and they are running off of the same amp so I speak from daily use and side by side experience of almost 2 years.

No one, and I mean no one- that has extensively listened to both will tell you that the CM5 is better.

It is possible that some people prefer the sound of the speakers that have less stellar measurements.

Objectively, was the comparison done “blind” with instant switching and level matching to remove as much bias as possible ?

 

[<eargiant]

It is possible that some people prefer the sound of the speakers that have less stellar measurements.

Objectively, was the comparison done “blind” with instant switching and level matching to remove as much bias as possible ?

Is it possible that a frequency response plot of a speaker is NOT the the be-all and end-all of a speakers measurements? Is it possible that it does not tell the FULL story?

Blind test them all you want, anyone with decent hearing and a good understanding/education of what they should be listening for (just as Harmon likes to prep subjects) will undoubtedly choose the 805D every time. How anyone can suggest otherwise is beyond me.

 

[PENG]

What do you have your R900 hooked up to and how loud do you usually have your volume? Also do the R900’s come with grills? I have a 2 year old where some kind of grill is a must to help against little fingers.

Most of the time I use the Parasound Halo A21 and Cambridge Audio 840a preamp. The A21 runs in class A for the first 8-10W and the CA is class A. I typically listen to spl of at least 10 dB below reference, the amp output would be fractional with peaks to no more than 25 W at any time. Those 2X8" drivers are very bass capable, much better than I expect based on their humble specs. The highs are not as transparent as my Focal 1028Be, but at half the price I am not complaining. For a more brilliance sound, I would probably give the 804 D2 (never heard the D3) the edge, but that's subjective, if you prefer the so called "warm" sound, you may actually prefer the R900. The reference 3 sounded closer to the B&W, but I am put off by the higher distortions in the low band below around 120 Hz.

They come with grilles and I find no reason to remove them. They are easily to remove and replace because they are held by hidden magnets.

For something on the neutral and accuracy side, aside from KEF, you may also check out Revels, such as the F208 that should be able to take the expected abuse better.

https://revelspeakers.com/productdetail/~/product/f208.html

Stereophile's JA seem to like it, but that's subjective, the measurements look great though.

"Summing up the Revel's measured performance is easy: In every way, this is textbook loudspeaker design. It's no wonder that Erick Lichte liked this speaker as much as he did.—John Atkinson"

Read more at https://www.stereophile.com/content/revel-performa3-f208-loudspeaker-measurements#dLvY6jgeDrU6vsK5.99

Another review with measurements:

https://hometheaterhifi.com/reviews/speaker/floor-standing/revel-f208-floor-standing-speakers-review/

I listened to them almost side by side with the KEF reference, ended up buying the R900.

 

[AcuDefTechGuy]

Is it possible that a frequency response plot of a speaker is NOT the the be-all and end-all of a speakers measurements? Is it possible that it does not tell the FULL story?

Blind test them all you want, anyone with decent hearing and a good understanding/education of what they should be listening for (just as Harmon likes to prep subjects) will undoubtedly choose the 805D every time. How anyone can suggest otherwise is beyond me.

It is possible.

At the same time, it is certainly possible that bias plays a much more significant factor than some people think.

Anything is certainly possible.

I won't get into an argument about blind-testing. I was just asking for anyone else who may also be interested in knowing.

So your answer is that no, your comparison was never blind-tested with level-matching and instant-switching. The comparison was done from pure memory knowing fully which speakers was which.

I just wanted to state the facts, not make an argument.

People can form their own opinions as they have the right.

If you were a friend or colleague, I would ask the same exact question.

Again, I am not trying to start an argument about blind-testing or anything.

 

[PENG]

I would say FR plots do not tell the whole story, but it is useful. Dr. Sean Olive has this to say about preference of neutral speakers:

"It seems that given some finite choices people will pick the most neutral speaker or headphone (no resonances), wide bandwidth. However, given some tone controls they will adjust for variations in program and taste."

http://seanolive.blogspot.ca/2015/11/factors-that-influence-listeners.html

Below are some FR measurements from Stereophile, the KEFs have better overall response. I believe distortion is also very important, and KEF R series and above seem to do well in this area too. B&W probably do well too but I couldn't find much data from the internet.

B&W 805 d3

B&W CM5

KEF LS50

KEF R700

 

[AcuDefTechGuy]

I would say FR plots do not tell the whole story, but it is useful. Dr. Sean Olive has this to say about preference of neutral speakers:

"It seems that given some finite choices people will pick the most neutral speaker or headphone (no resonances), wide bandwidth. However, given some tone controls they will adjust for variations in program and taste."

B&W 805 d3

Here is the S&V measurements of the B&W 805 D2 ("Diamond"):

https://www.soundandvision.com/content/bowers-wilkins-805-diamond-speaker-system-ht-labs-measures

You can see that B&W loves to spike the treble of their Diamond speakers.

This was what JA (Stereophile) was referring when he said that the treble was "tailored" to a certain sound.

It's the same treble spike with the 804D, 802D, and 800D.

https://www.stereophile.com/content/bowers-wilkins-802-d3-diamond-loudspeaker-measurements

https://www.stereophile.com/content/bampw-800-diamond-loudspeaker-measurements

https://www.stereophile.com/content/bowers-amp-wilkins-804-diamond-loudspeaker-measurements

 

[PENG]

Here is the S&V measurements of the B&W 805 D2 ("Diamond"):

https://www.soundandvision.com/content/bowers-wilkins-805-diamond-speaker-system-ht-labs-measures

You can see that B&W loves to spike the treble of their Diamond speakers.

This was what JA (Stereophile) was referring when he said that the treble was "tailored" to a certain sound.

It's the same treble spike with the 804D, 802D, and 800D.

https://www.stereophile.com/content/bowers-wilkins-802-d3-diamond-loudspeaker-measurements

https://www.stereophile.com/content/bampw-800-diamond-loudspeaker-measurements

https://www.stereophile.com/content/bowers-amp-wilkins-804-diamond-loudspeaker-measurements

Just 3 points I would like to make.

1. S&V's FR plots are not very useful due to the relatively low resolution and averaging (on axis, +/- 15 deg off axis) effect.

2. Stereophile's FR also seem lacking in resolution but at least they include the off axis graphs.

3. Stereophile also provided the CSD (cumulative spectral decay) plot, that apparently is a better for identifying resonances. My dispute with TheWarrior on the alleged resonance (by him) was partially based on that.

FR is obviously an important one, as Dr. Floyd Toole said in the AH Article:

"In a nutshell, if a speaker exhibits flat and linear on-axis frequency response with consistent off-axis performance to preserve critical early reflections, then the speaker will score very highly in blind listening tests and also provide more consistent performance from room to room."

https://www.audioholics.com/loudspeaker-design/measure-loudspeaker-performance

 

[AcuDefTechGuy]

Just 2 points I would like to make.

1. S&V's FR plots are not very useful due to the relatively low resolution and averaging (on axis, +/- 15 deg off axis) effect.

2. Stereophile's FR also seem lacking in resolution but at least they include the off axis graphs.

3. Stereophile also provide the CSD, that apparently is a better for identifying resonances. My dispute with TheWarrior on the alleged resonance (by him) was partially based on that.

FR is obviously an important one, as Dr. Floyd Toole said in the AH Article:

"In a nutshell, if a speaker exhibits flat and linear on-axis frequency response with consistent off-axis performance to preserve critical early reflections, then the speaker will score very highly in blind listening tests and also provide more consistent performance from room to room."

https://www.audioholics.com/loudspeaker-design/measure-loudspeaker-performance

Well, of course, measurements from Audioholics are better.

 

[yepimonfire]

Just 3 points I would like to make.

1. S&V's FR plots are not very useful due to the relatively low resolution and averaging (on axis, +/- 15 deg off axis) effect.

2. Stereophile's FR also seem lacking in resolution but at least they include the off axis graphs.

3. Stereophile also provide the CSD, that apparently is a better for identifying resonances. My dispute with TheWarrior on the alleged resonance (by him) was partially based on that.

FR is obviously an important one, as Dr. Floyd Toole said in the AH Article:

"In a nutshell, if a speaker exhibits flat and linear on-axis frequency response with consistent off-axis performance to preserve critical early reflections, then the speaker will score very highly in blind listening tests and also provide more consistent performance from room to room."

https://www.audioholics.com/loudspeaker-design/measure-loudspeaker-performance

This has been my experience a/b'ing speakers.

Sent from my LM-X210(G) using Tapatalk

 

[Dennis Murphy]

Here is the S&V measurements of the B&W 805 D2 ("Diamond"):

https://www.soundandvision.com/content/bowers-wilkins-805-diamond-speaker-system-ht-labs-measures

You can see that B&W loves to spike the treble of their Diamond speakers.

This was what JA (Stereophile) was referring when he said that the treble was "tailored" to a certain sound.

It's the same treble spike with the 804D, 802D, and 800D.

https://www.stereophile.com/content/bowers-wilkins-802-d3-diamond-loudspeaker-measurements

https://www.stereophile.com/content/bampw-800-diamond-loudspeaker-measurements

https://www.stereophile.com/content/bowers-amp-wilkins-804-diamond-loudspeaker-measurements

It would appear that the 10k spike is just a characteristic of the diamond tweeter rather than a deliberate voicing decision. They would need a trap circuit to deal with it, and those can be tricky in mass production due to their sensitivity to rather smaller variations in driver parameters. Of course, the bottom line is that there is still a big old peak at 10 k.

 

[shadyJ]

It would appear that the 10k spike is just a characteristic of the diamond tweeter rather than a deliberate voicing decision. They would need a trap circuit to deal with it, and those can be tricky in mass production due to their sensitivity to rather smaller variations in driver parameters. Of course, the bottom line is that there is still a big old peak at 10 k.

If that peak was a bit higher in frequency or if it was a higher Q peak, that wouldn't be much to worry about, but such a broad peak at 10kHz is bound to make the treble a bit sharp. Those speakers probably sound very detailed, but instrumentals with lots of high-frequency percussion like hi-hats and cymbals could end up being harsh.

 

[TLS Guy]

It would appear that the 10k spike is just a characteristic of the diamond tweeter rather than a deliberate voicing decision. They would need a trap circuit to deal with it, and those can be tricky in mass production due to their sensitivity to rather smaller variations in driver parameters. Of course, the bottom line is that there is still a big old peak at 10 k.

I agree with you Dennis, I don't think the peak is intentional. I did not measure it as high as others, but it is there.

This shows it better.

You can see there is something not right in that region. I suspect it is most likely diaphragm related.

I suspect they use an exotic diaphragm for marketing reasons. I'm suspicious of hard domes. I know B & W say it does not break up until an octave above human hearing. I have my suspicions that assertion is not entirely correct.

I know you like ribbons, but to me the SEAS excel Millenium tweeters are the best I have come across. Very flat with excellent dispersion, take a lot of power and sound excellent especially on the high strings.

That looks a lot better than the Diamond tweeter.

I suppose the critics would not be happy with a soft dome. However I think they sound the better and would take that slight high end edge of that Diamond series. They would have to come up with a new name though!

 

[Art Vandelay]

WRT to diamond dome, the response you see is the response expected.

All speaker drivers become more directional with frequency as the membrane accounts for > half a wavelength. That explains the rising response above about 6kHz. The response falls again beyond about 13kHz due to destructive interference effects - determined mostly by the dome geometry.

All 1 inch domes suffer the problem to some extent as per the laws of physics, but the simple remedy is to listen at a point that's off axis by about 20 degrees.

 

[AcuDefTechGuy]

If that peak was a bit higher in frequency or if it was a higher Q peak, that wouldn't be much to worry about, but such a broad peak at 10kHz is bound to make the treble a bit sharp. Those speakers probably sound very detailed, but instrumentals with lots of high-frequency percussion like hi-hats and cymbals could end up being harsh.

Also my sentiment when I owned the 802D2. I think they were a bit sharp and fatiguing with some materials after about an hour. But they look so darn cute.

 

[AcuDefTechGuy]

I suppose the critics would not be happy with a soft dome. However I think they sound the better and would take that slight high end edge of that Diamond series. They would have to come up with a new name though!

Well, it might measure much better, be less harsh, costs significantly less money, but it would certainly change that sound that some fans might have come to like.

The B&W Millennium Series: 800M, 802M, 803M, 804M, 805M.

But, what about Beryllium (Revel, Focal) or Titanium (KEF) ?

I don’t recall any big broad peak at 10KHz on the Salon2 or KEF 201/2.

The B&W Beryllium series?

B&W 800B, 802B, 803B, 804B, 805B.

But since the KEF 201/2 with its Titanium tweeter measured better than most speakers, perhaps it should be the B&W Titanium: 800T, 802T, 803T, 804T, 805T.

 

[PENG]

I agree with you Dennis, I don't think the peak is intentional. I did not measure it as high as others, but it is there.



This shows it better.

You can see there is something not right in that region. I suspect it is most likely diaphragm related.

I suspect they use an exotic diaphragm for marketing reasons. I'm suspicious of hard domes. I know B & W say it does not break up until an octave above human hearing. I have my suspicions that assertion is not entirely correct.

I know you like ribbons, but to me the SEAS excel Millenium tweeters are the best I have come across. Very flat with excellent dispersion, take a lot of power and sound excellent especially on the high strings.

That looks a lot better than the Diamond tweeter.

I suppose the critics would not be happy with a soft dome. However I think they sound the better and would take that slight high end edge of that Diamond series. They would have to come up with a new name though!

The second CDS looks better in the very high end but worse in the 2-6.5 kHz band. The Y-axis T represents "Period" right?

 

[PENG]

but the simple remedy is to listen at a point that's off axis by about 20 degrees.

Or apply EQ, eventually, even die hard audiophiles will embrace EQ like they do with digital source/player.

 

[AcuDefTechGuy]

It would appear that the 10k spike is just a characteristic of the diamond tweeter rather than a deliberate voicing decision. They would need a trap circuit to deal with it, and those can be tricky in mass production due to their sensitivity to rather smaller variations in driver parameters. Of course, the bottom line is that there is still a big old peak at 10 k.

Perhaps that's why Focal and KEF do not use Diamond tweeters in their $200K speakers.

Diamond tweeters appear to be just more exotic than actually good.

 

[Art Vandelay]

Or apply EQ, eventually, even die hard audiophiles will embrace EQ like they do with digital source/player.

Very true.

With room / speaker correction software available these days many already do. I've tested Dirac and it's pretty good IMO. You can dial up whatever response you like.

Fwiw, this is a Stereophile measurement of the Signature 800 which uses the old 1 inch aluminium dome. Note that the 10kHz bump is there exactly as it is with the diamond dome. The difference is the visibility of the dome resonance. So the Diamond is technically better, definitely, but whether it translates to something audible is debatable.