B&W 800D4 series

[MarcT]

Ugly as hell. But it's got a letterbox port, not many speakers can claim that.

Well, the Focal Alto Utopia Be that I'm looking around for a used set of has one!

 

[Pearljam5000]

Well, the Focal Alto Utopia Be that I'm looking around for a used set of has one!
View attachment 154680

They're still using it today

 

[Kal Rubinson]

Made me think of this:

 

[Music1969]

Just seen 801 D4 measurement in new HiFi News Mag

Very average FR, to put it lightly

These are made for people that want to show off and existing people upgrading that are happy with the house sound

Not made for well balanced / neutral playback

 

[preload]

I brought it up because you said this:

"People can stare at FR graphs all day and pretend that they have "a gift" and can reliably predict how speakers will sound based on eyeballing them (and btw, these aren't even full spinoramas that you're posting). But the reality is that Harman's own research demonstrates that computerized analysis of a full spin measurements can only explain 74% of the variability in listener preference scores." (Emphasis mine)

And what you said is true, but I also think the 74% figure isn't fully relevant to the question of whether measurements can determine what a speaker 'how speakers will sound.' Only whether they will be preferred (by a large sample of listeners in controlled conditions, of course). E.g. two people can hear the same sets of speakers, describe their sound similarly, and still have a difference of opinion on which is better.

Not sure if I agree, my sense is that it's even harder to pinpoint the sound qualities of a speaker (i.e. harsh, bright, forward, fatiguing, balanced, etc.) from an FR chart, than it is to simply determine if it will have high subjective sound quality overall (i.e. the "preference score"). This is especially true since eyeballing can give you a quick sense of how smooth the FR is and you can easily identify the bass extension. Whereas even to make a seemingly "easy" judgement like "the speaker will sound bright" (a common past time of the anti-B&W mafia), you would need to be able to differentiate between fractions of degree in the slope of a regression line, or decide what dB and Q peaks in the treble would be audible and the extent to which they would be perceived as "bright" relative to the "anchoring" levels in the midbass region.

So to be clear, I don't think I can definitively back up the notion that people can significantly figure out what a speaker will sound like based on measurements, but I also don't think you can cite existing research to suggest that we can't.

I wouldn't need to. The burden of evidence is on supporting the incredulous claim/ability.

To this point, speaking anecdotally with all the biases that entails... after 40+ speakers extensively measured and listened to, I generally find it quite easy to associate the measurements with a "what a speaker sounds like" for at least my own listening in my own home. Not perfectly, but to a high enough degree that I feel comfortable predicting sound just looking at extensive measurements, or guesstimating measurements from just listening. I recognize that might not translate as easily to other listeners, but whether because I'm fooling myself or it's the real truth of it, that's been my experience.

Great, and after listening to a loudspeaker, if you showed me the FR measurements, I'd be happy to "point out" all the FR characteristics that make it sound the way it does too. C'mon. That's like a palm reader saying "oh you won the lottery last night? Well that obviously happened because of this line on your palm right here."

I don't think we're that different in our opinions overall though. I do often think people here overestimate how pejorative some deviations from flat can be. In reality I think comparisons tend to be much 'fuzzier' than sometimes we in the measurement crowd think and the error bars are large.

Perhaps.

 

[ctrl]

Hmmm, I can't confidently argue with this data, but logically this f3 seems rather low. The 800D3 has two 10" drivers in a very large cabinet; 37Hz seems like a very strange f3 for a speaker this big. Without data of my own, I can't counter this data, but color me very skeptical. As a comparison, the Revel F208(2x8" drivers) has an f6 of 26Hz.

F208
Dimensions: HxWxD 1.2m x 0.34m x 0.38m.
The F208 has an f3=44Hz and f6=29Hz with a sensitivity of [email protected]. The BR tuning frequency is about 27Hz (see impedance plot).

Source: ASR

800D3
Dimensions: HxWxD 1.2m x 0.41m x 0.61m - about 1m height for the woofers.
Between 100 and 70Hz the 800D3 has a slight hump, a bass boost, which is at 92dB sound pressure level, but sensitivity is about [email protected].
As mentioned before, f3=37Hz and f6=29Hz (a second source, audio.de 04 /2017, measured the same f6=29Hz). The BR tuning frequency is about 26Hz (see impedance plot).

Source: fidelity-online.de

Both speakers have comparable bass, with different sensitivity. The 800D3 delivers 2-3dB more sound pressure level at the same voltage. Makes a certain sense, since the effective piston area of the 800D3 with quite roughly 700cm², clearly exceeds the approximately 440cm² of the F208.

The speaker cabinet of the 800D3 is not particularly large for two 10'' drivers with BR tuning.
The required cabinet volume of a BR speaker is proportional to the TSP parameter "Vas".
Vas = 0.0014 * Cms * Sd^2
with Cms=Compliance , Sd=eff. piston area

When doubling the eff. piston area, Vas quadruples.
If you now compare 800D3 with F208 and compare the eff. piston area, the volume difference fits quite well.

If we now say in a very simplified (not quite correct) way that the low bass response of a driver is coupled to the free air resonance (fs) of the driver, the lower fs, the lower the BR tuning can be done.
It applies to fs:
fs = 1 /2 * pi * (Cms * Mms)^0.5
with Mms=Moving mass of the driver, Cms=Compliance

This makes it clear that a large loudspeaker does not necessarily mean a deep bass response (no volume dependence) and that, conversely, a small driver with the appropriate "moving mass" (call it subwoofer driver ) can play very low (with low sensitivity).

 

[Matias]

Just seen 801 D4 measurement in new HiFi News Mag

Very average FR, to put it lightly

These are made for people that want to show off and existing people upgrading that are happy with the house sound

Not made for well balanced / neutral playback

Somewhat tilting to treble but flatter than 802D3 on HFN 12/2015, which has an upper bass and low mids valley. But the D4 flagship had a great subjective review and grade FWIW.

 

[Deleted member 27948]

Could imagine a f3 (frequency response is -3dB below the average sound pressure level of the loudspeaker) of 22-30 Hz as realistic, depending on whether the [email protected] sensitivity apply. ...
Independent measurements then confirmed the sensitivity of 90dB (even reported 90.7dB), but for that, the low-frequency capability claims were pure fantasy. An f3 of about 37Hz and an f6 of 29Hz are galaxies away from the manufacturer's claims of 15Hz.
View attachment 154656

Update: When someone spends $30,000 or more on a speaker, it's probably to give the buyer the feeling that Bon Scott so aptly put it, "But we've got the biggest balls of them all."

My 2c: The LF FR response above is, obviously, from a truncated IR, measured by the "semi-anechoic" technique. You may ask why "obviously"? Sorry, I am not capable of explaining that to non-PhDs in a forum post.

Generally, measuring such low-frequency as 20Hz requires open-field conditions in a sphere of > 15m diameter. Such amenities are not common but do exist, and can be rented for a relatively low price. For example, there is a "chamber" near Seattle made from an old ... let's say, airplane, construction space. The port forms a 4th order IIR, which has poles quite close to the unity circle, and you'll need to either catch 99% of energy in the IR tail, which is 15m*(4...5). Alternatively, you may apply modern advanced parametric methods instead of default stone-aged non-parametric periodogram. Reviewers rarely have such knowledge, if they do, they would not work as reviewers but as Chief/Sr. Scientists.

The suitable drivers are not uncommon, see https://www.focal-america.com/wp-content/uploads/2014/02/FT_access_25A4.pdf Put it in a 100+L vented box with a port tuned, say, at 20Hz, and voila!

But I do agree with your note about balls. Technically, 804D4 + a decent closed box sub are way cheaper than 801.

 

[preload]

My 2c: The LF FR response above is, obviously, from a truncated IR, measured by the "semi-anechoic" technique. You may ask why "obviously"? Sorry, I am not capable of explaining that to non-PhDs in a forum post.

Most people on this forum have self-awarded Ph.D.'s in spinorama interpretation.

 

[Beave]

My 2c: The LF FR response above is, obviously, from a truncated IR, measured by the "semi-anechoic" technique. You may ask why "obviously"? Sorry, I am not capable of explaining that to non-PhDs in a forum post.

My guess was that the freq response plot was a splice of a close-mic measurement for the port & woofers and a pseudo-anechoic "gated" measurement for the mid-tweeter combo.

Generally, measuring such low-frequency as 20Hz requires open-field conditions in a sphere of > 15m diameter. Such amenities are not common but do exist, and can be rented for a relatively low price

. For example, there is a "chamber" near Seattle made from an old ... let's say, airplane, construction space. The port forms a 4th order IIR, which has poles quite close to the unity circle, and you'll need to either catch 99% of energy in the IR tail, which is 15m*(4...5). Alternatively, you may apply modern advanced parametric methods instead of default stone-aged non-parametric periodogram. Reviewers rarely have such knowledge, if they do, they would not work as reviewers but as Chief/Sr. Scientists.

There are lots of people measuring subwoofers using ground-plane measurements. Are they all wrong? Their data might not be *quite* as good as what could be achieved in a very large open space, but it's probably good enough for home audio purposes, right?

 

[Deleted member 27948]

There are lots of people measuring subwoofers using ground-plane measurements. Are they all wrong? Their data might not be *quite* as good as what could be achieved in a very large open space, but it's probably good enough for home audio purposes, right?

Rhetoric questions aside... how many of them know the theory of spectral analysis, delta-autocorrelated and perfect sequences, M-sequences, exponential chirp function, system estimation theory, acoustic theory, etc? do they understand what they are doing? do they know what problems they can run into? can they spot a bug in their measurements? do they know how to fix it? can they scientifically assess the precision of their measurements, validate gaussianity, compute Fisher information matrix, etc? Thus, shall anybody trust their measurements... most likely NO.

do they have fun? most likely YES!

 

[Beave]

Are any of those needed for measuring subwoofer performance? Most likely NO, to borrow your words.

This is home audio, not rocket surgery*.


*Intentional malapropism

 

[preload]

The room AND the placement of the speakers in the room are going to be far more important than how the FR was measured to determine the F3.

Also, why F3? Olive's regression formula used the F6 of the sound power curve, and I'm pretty sure it was intentional.

 

[Deleted member 27948]

Are any of those needed for measuring subwoofer performance? Most likely NO, to borrow your words.

This is home audio, not rocket surgery*.


*Intentional malapropism

I expect many will disagree with me, which I don't mind.

That depends on the definition of the word "measure": if it means "measure", then YES. if it means "cargo cult", then NO. you can't use math properly without understanding it, same as knowing words karate / sambo / box / etc may help you look cool while talking to other teenagers but does not turn you into a martial arts master:-(

 

[MediumRare]

I expect many will disagree with me, which I don't mind.

That depends on the definition of the word "measure": if it means "measure", then YES. if it means "cargo cult", then NO. you can't use math properly without understanding it, same as knowing words karate / sambo / box / etc may help you look cool while talking to other teenagers but does not turn you into a martial arts master:-(

Someone once explained to me that F1 drivers rely on an exceptional understanding of vector physics for, not only their success, but their lives. They do not, however, do calculations.

 

[richard12511]

Are any of those needed for measuring subwoofer performance? Most likely NO, to borrow your words.

This is home audio, not rocket surgery*.


*Intentional malapropism

Indeed, ground plane measurements can be very accurate (for the bass), more accurate than an anechoic chamber. Hartman even uses them for that purpose. Judging by Erin’s ground plane measurement pictures, he definitely has at least 15m+(likely much more) on all sides.

Hopefully Erin ends up an 801D4 to measure, so we can see exactly how accurate the spec is.

 

[MarcT]

Someone once explained to me that F1 drivers rely on an exceptional understanding of vector physics for, not only their success, but their lives. They do not, however, do calculations.

Lol, it's rather tough to do any calculations in the middle of a turn at 150 mph! I think they just go by feel at that point.

 

[Vear]

Someone once explained to me that F1 drivers rely on an exceptional understanding of vector physics for, not only their success, but their lives. They do not, however, do calculations.

Because the F1 drivers exceptional understanding is subjective based on how they feel in the motorcar as it navigates the course.

 

[richard12511]

The room AND the placement of the speakers in the room are going to be far more important than how the FR was measured to determine the F3.

Also, why F3? Olive's regression formula used the F6 of the sound power curve, and I'm pretty sure it was intentional.

I think the reason we’re talking about F3 is because that’s what B&W chose to use in the spec.

 

[richard12511]

F208
Dimensions: HxWxD 1.2m x 0.34m x 0.38m.
The F208 has an f3=44Hz and f6=29Hz with a sensitivity of [email protected]. The BR tuning frequency is about 27Hz (see impedance plot).
View attachment 154691
Source: ASR

800D3
Dimensions: HxWxD 1.2m x 0.41m x 0.61m - about 1m height for the woofers.
Between 100 and 70Hz the 800D3 has a slight hump, a bass boost, which is at 92dB sound pressure level, but sensitivity is about [email protected].
As mentioned before, f3=37Hz and f6=29Hz (a second source, audio.de 04 /2017, measured the same f6=29Hz). The BR tuning frequency is about 26Hz (see impedance plot).
View attachment 154695
Source: fidelity-online.de

Both speakers have comparable bass, with different sensitivity. The 800D3 delivers 2-3dB more sound pressure level at the same voltage. Makes a certain sense, since the effective piston area of the 800D3 with quite roughly 700cm², clearly exceeds the approximately 440cm² of the F208.

The speaker cabinet of the 800D3 is not particularly large for two 10'' drivers with BR tuning.
The required cabinet volume of a BR speaker is proportional to the TSP parameter "Vas".
Vas = 0.0014 * Cms * Sd^2
with Cms=Compliance , Sd=eff. piston area

When doubling the eff. piston area, Vas quadruples.
If you now compare 800D3 with F208 and compare the eff. piston area, the volume difference fits quite well.

If we now say in a very simplified (not quite correct) way that the low bass response of a driver is coupled to the free air resonance (fs) of the driver, the lower fs, the lower the BR tuning can be done.
It applies to fs:
fs = 1 /2 * pi * (Cms * Mms)^0.5
with Mms=Moving mass of the driver, Cms=Compliance

This makes it clear that a large loudspeaker does not necessarily mean a deep bass response (no volume dependence) and that, conversely, a small driver with the appropriate "moving mass" (call it subwoofer driver ) can play very low (with low sensitivity).

Thanks for the detailed explanation. A really good comparison would be the Revel F328Be. I has 3 8" drivers, with a very large enclosure and identical sensitivity. It has an f3 of 35Hz.