New Bowers and Wilkins 800 series...

[Patrick Butler]

Hi Bob,

 

The casters were only intended to facilitate moving the speakers into position. They negatively impact performance, and we have always recommended that spikes be used. The pointy end should be used with carpeted surfaces, and the rubber end with wood, tile or stone surfaces.

 

When adjusting rake (tilting forward or back in the vertical plane) you should always start with making the speakers plumb using a level. I used a standard level that has multiple gauges so it can be used vertically or horizontally. The front baffle can be used to adjust rake forward and back, and the side of the speaker (bass cabinet) can be used for adjusting side to side. Once both channels are plumb, then you can being adjusting them for a given listening height and distance.

 

Imagine a sphere projected 9'10" in front of each speaker. Localized within that sphere is the optimal alignment of all drivers. What you are in essence trying to do is to adjust each channel independently so that both spheres perfectly focus in a listening area. With this in mind, it is time to start listening. Throw on a recording of a woman singing with a bass accompaniment if possible. As you sit in your listening chair move a little to the left and focus on that channel. Move your head slowly up so that you begin to listen above the tweeters axis, and then down through that axis until you are a bit below that midrange axis. You'll hear an area where there is a perfect balance between mid and upper frequencies. You'll notice that it sounds both clear and natural (a balance of upper to mid frequencies.

 

Once you locate that spot, you can adjust the front spikes to shift that area to a height that corresponds with your head. You won't need a large adjustment, so go slowly. Once you've taken care of the left channel, all you need to do is adjust the right channel to match. Also pay attention to the bass. If bass quality is negatively impacted, you might have gone a bit too far in your adjustment. Simply back off a bit and you should find that bass quality returns, and the critical alignment of all other drivers is achieved.

 

That is the simple explanation.

 

Regards,

 

Patrick

B&W Group North America

 

Well, at least I got the listening distance right!

 

Patrick, do you have any tips on exactly how to adjust the speaker in the vertical plane? How do you choose a specific amount of downward/forward tilt to compensate for lower than optimum listening height? I purchased higher chairs but I am afraid that the increase in chair height was negated by the increase in speaker height when I put the feet on. I couldn't get the speakers level on my tile floor with the castors as one or two castors were always sitting in the groove between adjacent tiles, therefore I had to resort to using the feet.

 

Cheers,

Bob

[Sam Lord]

Hi Sam,

 

Apologies for the late reply. Been on the road this week with long days and late nights.

 

The minimum listening distance for 802D2 or 802D3 would be 9' 10". That is where all the drivers align. You can listen further away if wished (and some definitely do,) but considerably closer will result in the drivers not aligning to their full potential.

 

 

Regards,

 

Patrick

B&W Group North America

 

Thank you Patrick! You are awarded the *first* Gold Star! I am *completely* satisfied with your answer and I thank you. Feel free to ignore what follows.

_____

 

I do suggest you add this to the manual. I am curious what you folks think about the degree of audible tolerance for phase error, and whether it changes with frequency. In my experience, it didn't. For the mid-tweeter alignment where we used a 1st-order crossover at 3kHz, we translated the tweeter fore or aft by increments of <1/16". Our bass-to-midrange offsets, with an xover point around 400Hz, produced audible changes at about 0.3". The meaning of the acoustic center between two drivers, and the shape of drivers obviously matters, but we found that when one of us thought the sound had really, finally snapped just right at a listening position, everyone agreed about it.

 

I don't know whether you build crossoverless versions of your speakers by request. I would truly love to use a digital crossover with great precision for the slopes, perhaps something that Jussi Laako would write (or has written). I would keep the slopes at something like 24dB/octave, only without the 360deg lag of of LR passive designs, easy to calculate into values. If I did any room correction at all it would gently and simply address only the dominant pole in the room. In my dreams, understand. I think B&W makes the best bass drivers in the industry, and the mid and high transducers are right up there as well.

 

...Imagine a sphere projected 9'10" in front of each speaker. Localized within that sphere is the optimal alignment of all drivers. What you are in essence trying to do is to adjust each channel independently so that both spheres perfectly focus in a listening area. With this in mind, it is time to start listening...

 

That was an excellent response to Bob's question; I strongly request that we make it a sticky.

[Axiom05]

Thanks, Patrick. I will see what I can do with your detailed information.

 

Cheers!

[徐中銳]

Martial Rousseau, listed as « Head of research at B&W Group » co-presenting with Marantz' key « Quality Control manager » Ryuichi Sawada at the Tokyo International Audio Show 2015 :

 

And, how Marantz Japan have been B&W's official importer since 1993, below is Sawada's positioning of the new and old models at his listening room :

 

 

 

[MikeJazz]

I don't know whether you build crossoverless versions of your speakers by request. I would truly love to use a digital crossover with great precision for the slopes, perhaps something that Jussi Laako would write (or has written). I would keep the slopes at something like 24dB/octave, only without the 360deg lag of of LR passive designs, easy to calculate into values. If I did any room correction at all it would gently and simply address only the dominant pole in the room. In my dreams, understand.

 

This might be of interest to you. Exakt technology for 802D's.

Get Linn Exakt into a B&W 802D | The Ear

 

B&W 802 Diamond Exakt conversion kit

[sharbatgula]

Wow, I stumbled upon a great forum full of knowledgeable people and plenty of useful information!

I apologize for not hitting the target 100% about the topic (800 D series), but the observations I post are related to the observations that some of you discuss.

It's the same with all speakers in the Diamond range. There's a 3+ dB hump.

In the corner of the world where I live, I never had a chance to audition a Diamond range of speakers, but I find that hump present across the 600 S1 and S2 and the CM series. Also measurements from many respectable other vendor's speakers show that hump more or less. I always wondered why? My measurements of my 685 S1 show a huge hump in that range and bar for the female voices I dislike that behaviour as piles of other music sounds out of balance to me. And from there started my quest for speakers that have the positives of B&W, but without the invasive lower mids.

Enter B&W PM1, the only B&W speaker (that I could audition) behaving differently is the now discontinued PM1.

From what I auditioned and the very hard to find measurements (Stereoplay, Kompakt-Lautsprecher B&W PM1 im Test - connect), this speaker is much less forward, but still detailed and nicely balanced. There is small hump at about 1.2 kHz, followed by a dip at about 1.8 kHz and from then on the energy in the highs is rising, giving it a feeling of airiness and much welcome elevated highs for my aging ears.

 

The Kevlar mid's only vice is a small break-up mode at 3.5kHz.

 

Do you think that the smaller size of the PM1 midbass cone (4.5" without the surround) sufficiently moves the breakup much higher above the crossover point (4 kHz, simple cap and coil design)?

 

Do you think that the smaller midbass size also contributes to larger sweet spot, i.e. less lobing problems, lesser sensitivity to vertical positioning of the ears? It's hard to find reviews of PM1 and people on forums report excellent integration of the units, despite the first electrical order crossover.

 

My preference is musical, neutral sound while still preserving detail, but not analytical and detail for the sake of it. I like to listen for hours, often off the hot spot.

PM1 sounded pretty good to me and I'm inclined to pull the trigger and buy the only remaining pair. But it bears risks as for my circumstances it is still very expensive and home auditioning is not allowed. I'm afraid that in-house, kevlar and the high crossover frequency could still show the bad face.

Any experience with PM1s. Am I safer with the smaller cone size?

 

Thank you!

[RFP]

 

In the corner of the world where I live, I never had a chance to audition a Diamond range of speakers, but I find that hump present across the 600 S1 and S2 and the CM series. Also measurements from many respectable other vendor's speakers show that hump more or less. I always wondered why? My measurements of my 685 S1 show a huge hump in that range and bar for the female voices I dislike that behaviour as piles of other music sounds out of balance to me. And from there started my quest for speakers that have the positives of B&W, but without the invasive lower mids.

Enter B&W PM1, the only B&W speaker (that I could audition) behaving differently is the now discontinued PM1.

From what I auditioned and the very hard to find measurements (Stereoplay, Kompakt-Lautsprecher B&W PM1 im Test - connect), this speaker is much less forward, but still detailed and nicely balanced. There is small hump at about 1.2 kHz, followed by a dip at about 1.8 kHz and from then on the energy in the highs is rising, giving it a feeling of airiness and much welcome elevated highs for my aging ears.

 

Wrt the 800 series, the peak at 3-4kHz is partly a result of (controlled) break-up and partly a side effect of the 2nd order crossover with mechanical phase alignment. Fwiw, it also compensates for the 3dB dip in the power response at the same (crossover) frequency, so without it the speaker will sound slightly 'lacking' in this region in most environments. B&W design engineers are very competent, so this is not an accidental design move.

 

However, the perceived end-balance will also depend significantly on the listening distance, room size, furnishing etc, and can be modified further by adjusting the listening axis. The peak will be quite audible if listening in the near-field for instance, as would be more the norm with mini monitors, which is also why mini monitors are typically designed 'flat' on-axis while most large floor-standers are not designed 'flat'.

 

Do you think that the smaller size of the PM1 midbass cone (4.5" without the surround) sufficiently moves the breakup much higher above the crossover point (4 kHz, simple cap and coil design)?

 

I haven't seen as review of this speaker but yes, what you're suggesting is most likely correct. Given that this speaker is a stand mount 2-way design then I would expect that it would exhibit a flat on-axis response through the mid band.

 

Do you think that the smaller midbass size also contributes to larger sweet spot, i.e. less lobing problems, lesser sensitivity to vertical positioning of the ears? It's hard to find reviews of PM1 and people on forums report excellent integration of the units, despite the first electrical order crossover.

 

My preference is musical, neutral sound while still preserving detail, but not analytical and detail for the sake of it. I like to listen for hours, often off the hot spot.

PM1 sounded pretty good to me and I'm inclined to pull the trigger and buy the only remaining pair. But it bears risks as for my circumstances it is still very expensive and home auditioning is not allowed. I'm afraid that in-house, kevlar and the high crossover frequency could still show the bad face.

Any experience with PM1s. Am I safer with the smaller cone size?

 

Thank you!

 

If the PM1 employs low-order filters then the listening height is likely to be critical, but the smaller driver will also 'lobe' less at the top of the pass band, so it's likely to exhibit a reasonably flat power response to satisfy casual listening around the house as well as critical listening from the arm chair.

 

IME, the kevlar break-up is not a serious issue, even though the associated response peaks often look menacing in published reviews. Fortunately the waterfall decay is still quite clean, and the Q is high enough to make it reasonably inaudible, even when listening on-axis and at distances of less than 2m. The house sound to which you refer has more to do with the falling power response of a 6" driver in the 1-3kHz region, slightly exacerbated by the crossover design itself. Some refer to this as the "BBC dip", and FME it's a signature that's synonymous with many British loudspeakers.

 

If you're hearing a defined 'peak' with your 685's it's possible that the break-up is less well controlled than it is with the FST 800 series drivers, which would not be unexpected, but perhaps your listening room and distance is also contributing. With some experimenting of listening axis, distance, and height, you should still be able to achieve a near perfect tonal balance.

[sharbatgula]

Thank you RFP.

B&W design engineers are very competent, so this is not an accidental design move.

I have no doubt about the competence of the B&W engineers, but with the crossover frequency so high, I don't understand what is the gain, other than maybe cost savings (which would be miserable if true for such high-end speakers) of employing first order networks? Is it really that better sonically to justify a tiny hot spot?

[Axiom05]

Thank you RFP.

 

I have no doubt about the competence of the B&W engineers, but with the crossover frequency so high, I don't understand what is the gain, other than maybe cost savings (which would be miserable if true for such high-end speakers) of employing first order networks? Is it really that better sonically to justify a tiny hot spot?

 

I used to wonder about this as well, re. crossover frequency. I assumed that it was to increase power output in that frequency range, i.e., the FST driver could play louder and cleaner than the tweeter could at the bottom of its operating range. I'm usually wrong, however. I stopped worrying and now just listen and enjoy. When I purchased my 802 Diamonds2, they simply sounded the best of all the speakers that I auditioned, some costing more, some less. I didn't expect to buy another pair of B&W's, thought I was ready for a change, but the proof was in the listening.

[sharbatgula]

Another mistery that comes to my mind is why on so many two ways including the 600, CM and partly even PM1 I notice that as notes on pianos start to reach the fifth octave (or second if counted from the first in the middle of the piano), the bottom end of the tone crushes and those notes in that octave sound as if coming from an instrument with tiny box, not a large soundboard. It's hard for me to find the proper words, but the sound becomes thin and tinkly, totaly unconvincing compared ro real piano in a room. As the notes depart away up or down the sound becomes much more convincing.

I don't have an idea if this has something to do with the midrange hump, but the only exceptions from this are some dome midrange based speakers and a couple of ribons/air motion two ways, but never two ways with cone mid od midbass. The best of them in piano reproduction was a Gauder Arcona. What could be source of this problem? Is it something inherent to cone shape, breakup (too low frequency to be true), low order crossover or maybe phase distortion?

Anyone experienced the same thing?

[徐中銳]

Sound performance, on driving B&Ws, I recall Yasuhiro Sarashina explaining « momentary current » from a 2003 interview, might as well be in full :

What was the first product you worked on when you joined Marantz?

I joined Marantz in 1991, and the first product I worked on was the DCC (Digital Compact Cassette). As far as amplifiers go, I started working on them in 1995, with the Model 7, Model 8B and Model 9 replicas.

 

What have you worked on recently?

In terms of major projects, the SC-7S1 & MA-9S1.

 

So between 1991 and now, you saw two large developments with the start of Japanese distribution for B&W and the intoduction of Super Audio CD. In light of those developments, where do you start when developing Marantz' dedicated flagship SC-7S1 & MA-9S1 combination?

It's true that the introduction of Super Audio CD and the B&W speakers, especially the Nautilus 800 Series made life more difficult for the amplifier engineer. The B&W Signature Series especially are really hard to drive properly, so we started by looking into the best way to drive them.

When you try to playback a Super Audio CD using such hard-to-drive speakers, what kind of performance is required from the amp?

First of all, it has to have a very high signal-to noise ratio and have extended high frequency performance. Super Audio CD reaches to about 100 kHz in the high-end, so you need to secure a good amount of channel separation that extends to very high frequencies. Also, in addition to regular current supply, one has to look at how much current can be supplied quickly at any given moment. This was a completely new concept for us, and involves looking at how much current can be supplied to the speaker at very short notice, basically 1/1000 of a second. We measured this by sending a single cycle 1 kHz tone burst signal under a 0.1 Ohm load and measure how much fluctuation is achieved and work out the momentary current performance from that.

So, am I right in understanding that with previous current performances, you were looking at how much current was continuously available, whereas with momentary current performance, it's about looking at how much current can be supplied in any given moment?

When you do these tests with different amps that have the same power and low load power readings, they show different readings when it comes to momentary current. For example, the Marantz SM-5 and MA-9S1 show different performance characteristics in that respect.

 

Does achieving that kind of performance put weight on the design of the output stage and power supply?

In order to draw additional current at any given moment, the power transistors must not fail. Within Marantz we have strict guidelines to make sure that the transistors do not fail when a short occurs, even when the amplifier is operating at maximum capacity. To achieve that guideline, we have traditionally used current limiters on the power transistors, but with our new model, the MA-9S1, we opted to take out the current limiter and instead, monitor the current reading of every power transistor and use relays to prevent power transistor failure when trouble occurs. We are also using new transistors in the pre-driver stage that allow more current to pass through.

 

Traditionally, "Monster" amplifers are regarded as having enough capacity to power speakers, but having slow response. From that perspective, how does the MA-9S1 perform?

With the MA-9S1, we went to great lengths to make sure that it worked at its best as a mono amplifier under a bridged connection. We also made sure that the hot and cold signals were lined parallel, so that noises didn't 'jump out', and designed the circuit board layout so that the high current signal path length was kept as short as possible. Imagine two amplifiers making a bridged connection laid out in parallel on a single large heat sink, and you have the basics of the MA-9S1.

 

So you strived for a minimal signal path?

We made things basic, simple, and true to what amplifier design should be.

 

Do you draw the circuit board patterns on Computer using CAD software?

No, I do my layouts on graph paper and let a dedicated CAD operator input the data for manufacturing.

 

So, you need more experience with CAD?

Well, for me, my workings on paper are drafts that are used till completion, and CAD is merely a way to make a precise clean copy (laughs). All the circuits are eventually drawn on CAD, but you need to scroll through the screen to see the whole layout. Its hard to see the overall perspective and image of a design when you have to scroll through a screen. So, in my case, I tend to work 'analog' and design things on paper manually, this burns the circuit designs and layouts into my brain.

[Sam Lord]

This might be of interest to you. Exakt technology for 802D's.

Get Linn Exakt into a B&W 802D | The Ear

 

B&W 802 Diamond Exakt conversion kit

 

Thanks Mike, that's really useful.

 

Sound performance' date=' on [i']driving[/i] B&Ws, I recall Yasuhiro Sarashina explaining « momentary current » from a 2003 interview, might as well be in full :

 

This is exactly the argument Shunyata has made from its inception, and their power devices at least help the amps do the job.

[徐中銳]

Pardon Sam,

busy, much to do...

 

My records indicates that Sawada, 2 years ago, was using the (upper caseless) NSIT-2000Plus (now Mark II)—but only for the SA-7S1 Player and SC-7S2 Pre-amp, whereas « Power amps are connected directly from the electrical outlet » (original Japanese being) :

SA-7S1とSC-7S2はノイズ対策を兼ねて中村製作所のNSIT-2000Plusを経由して給電している。パワーアンプはコンセントからダイレクトにつながれている

Marantz Japan have been B&W's official importer since 1993' date=' below is of [b']Sawada[/b]'s listening room :

 

When I read negative performance-observations about B&Ws (or any other brands' loudspeakers) I wonder what amps were driving...

Marantz is but an example (and perhaps useful for comparisons while in shops that has their Amps too—when we're auditioning B&Ws)

[semente]

When I read negative performance-observations about B&Ws (or any other brands' loudspeakers) I wonder what amps were driving...

Marantz is but an example (and perhaps useful for comparisons while in shops that has their Amps too—when we're auditioning B&Ws)

 

I get the impression, both from listening and by looking at measurements, that BnW's newer designs don't perform was well as the old 801 Matrix models; frequency response (responsible for the overall tonal balance) is more ragged than and the dispersion characteristics aren't as even.

Even the spectral-decay looks better in S'phile's measurements of 801 Matrix 2.

 

As far as I understand, they have, commendably, been trying to improve the performance of their drivers and enclosures but I wonder if this ins't in someway related to the deterioration in technical performance.

Alternatively, the jagged upper-midrange/lower-treble response could be an intentional option to create some sort of "house" sound, since this raggedness is generally perceived subjectively as "lively" and "detailed".

 

In my opinion, BnW's kevlar midrange drivers on the speaker's I've listened to produce a noticeable, unpleasant, "coloration" at the top of their passband. It will be interesting to see if, with the new cone material, they've gotten rid of it (even if the raggedness of the upper-midrange does make it more difficult to identify)...

 

R

[Axiom05]

I get the impression, both from listening and by looking at measurements, that BnW's newer designs don't perform was well as the old 801 Matrix models; frequency response (responsible for the overall tonal balance) is more ragged than and the dispersion characteristics aren't as even.

Even the spectral-decay looks better in S'phile's measurements of 801 Matrix 2.

 

As far as I understand, they have, commendably, been trying to improve the performance of their drivers and enclosures but I wonder if this ins't in someway related to the deterioration in technical performance.

Alternatively, the jagged upper-midrange/lower-treble response could be an intentional option to create some sort of "house" sound, since this raggedness is generally perceived subjectively as "lively" and "detailed".

 

In my opinion, BnW's kevlar midrange drivers on the speaker's I've listened to produce a noticeable, unpleasant, "coloration" at the top of their passband. It will be interesting to see if, with the new cone material, they've gotten rid of it (even if the raggedness of the upper-midrange does make it more difficult to identify)...

 

R

 

Having owned the Matrix 801, Nautilus 801 and now the 802 Diamonds2, I can assure you that the sound has improved significantly despite what you think the measurements show. I was also concerned about the measured performance of the Diamond series as shown in Stereophile reviews, however the true test is what you hear. I actually chose the 802 Diamonds over the Magico S3 speakers, which was totally unexpected. B&W can probably make a speaker measure anyway they want, therefore there must be a reason they have settled on a specific design. My 802 Diamonds will not be leaving my listening room for a very long time. I have not yet heard the 802 D3. Forget your impression, go listen to some properly set up speakers and let your ears decide. You seem to make negative comments about the 800 Series whenever the opportunity presents itself. My impression is that you have not actually seriously auditioned any of these speakers.

 

Below is a link to a non-Stereophile review.

 

http://www.ultrahighendreview.com/uploads/documents/b&w_802diamond.pdf

[Totsipaki]

If my budget allowed (which it doesn't) that's the model I'd go for

 

Me too but I also feel lucky that I have managed to obtain my 803s a few years back..

[semente]

Having owned the Matrix 801, Nautilus 801 and now the 802 Diamonds2, I can assure you that the sound has improved significantly despite what you think the measurements show. I was also concerned about the measured performance of the Diamond series as shown in Stereophile reviews, however the true test is what you hear. I actually chose the 802 Diamonds over the Magico S3 speakers, which was totally unexpected. B&W can probably make a speaker measure anyway they want, therefore there must be a reason they have settled on a specific design. My 802 Diamonds will not be leaving my listening room for a very long time. I have not yet heard the 802 D3. Forget your impression, go listen to some properly set up speakers and let your ears decide. You seem to make negative comments about the 800 Series whenever the opportunity presents itself. My impression is that you have not actually seriously auditioned any of these speakers.

 

 

Thank you for the review.

I hardly ever read the subjective impressions but it was nice to learn that another magazine is performing measurements.

 

On fig.5 you can see the midrange driver's resonance peak at around 3.3KHz (green trace), right where the ear is most sensitive.

It's 5dB in amplitude and the effects on the overall frequency response are visible in both fig.1 and fig.5 (pink trace); these measurements correlate very accurately to what I have listened to at dealers and shows.

The tweeter also appears to be at least some 4dB too "hot".

 

 

As I have mentioned previously, BnW may have (probably did) make improvements both in driver and enclosure performance.

But the frequency response is far from what one would expect, at any price.

 

You've mentioned that "B&W can probably make a speaker measure anyway they want, therefore there must be a reason they have settled on a specific design"; so why don't they make it measure better?

Maybe Patrick Butler can answer that question.

 

 

As for my comments, they're just another opinion which happens not to match yours.

 

R

[徐中銳]

Okay, as with Sarashina explaining :

So between 1991 and now, you saw two large developments with the start of Japanese distribution for B&W and the intoduction of Super Audio CD. In light of those developments, where do you start when developing Marantz' dedicated flagship SC-7S1 & MA-9S1 combination?

It's true that the introduction of Super Audio CD and the B&W speakers, especially the Nautilus 800 Series made life more difficult for the amplifier engineer. The B&W Signature Series especially are really hard to drive properly, so we started by looking into the best way to drive them.

 

 

When you try to playback a Super Audio CD using such hard-to-drive speakers, what kind of performance is required from the amp?

First of all, it has to have a very high signal-to noise ratio and have extended high frequency performance. Super Audio CD reaches to about 100 kHz in the high-end, so you need to secure a good amount of channel separation that extends to very high frequencies. Also, in addition to regular current supply, one has to look at how much current can be supplied quickly at any given moment. This was a completely new concept for us, and involves looking at how much current can be supplied to the speaker at very short notice, basically 1/1000 of a second. We measured this by sending a single cycle 1 kHz tone burst signal under a 0.1 Ohm load and measure how much fluctuation is achieved and work out the momentary current performance from that.

Then, let's ask whether the reviewers' measuring device drives B&Ws properly for their testing-snapshot?

 

Additionally, it's true that many recording facilities monitor with B&Ws ? I'll quickly recall 2 I've spoke of before in CA, Polyhymnia : "Phillips Classics legacy" and Abby Road Studios' latest surround re-recording :

 

Thus, should those keen to replicate exactly what the mastering engineers hear use...

[semente]

I have quite a few recordings of "classical" music and the large majority sounds "wrong", both from a tonal balance and from an imaging perspective.

If the engineer cannot get the sound of recordings "right", I see no reason to trust their judgement when it comes to selecting speakers.

 

And, in extremis, studio monitors serve a different purpose than what's expected of speakers for domestic reproduction of recorded music: the former must present a "clear" window of the mic feed, EQ and mix, while the latter must sound "good" when reproducing recordings.

The same could be said for speakers used for testing/reviewing audio equipment.

 

R

[Patrick Butler]

Hi semente,

 

Great question! I guess the question would be measure better how? Total harmonic distortion, intermodulation distortion, thermal distortion of the voice coil, distortion caused by internal reflections in the cabinet, distortion caused by diffraction, distortion caused by the resonant frequency of the basket, etc, etc etc. Frequency response is but a single measurement, and where you put the microphone will determine how closely your measurements reflect a product's actual reference response. I often shake my head at some of the measurements that we see published in magazines. As an example, there is no peak in the tweeter's response at 3.5kHz or a dip around 5.5kHz or a peak around 8kHz. Just doesn't exist in the design.

 

Is a flat response the most important goal for our engineers? Categorically not. Important? Yes- but not the most important. The aforementioned 800 Series Matrix measured flatter. However, today's 800 Series (or the last generation for that matter) are much quieter designs with overall lower levels of distortion that sound hugely improved. At the end of the day, all loudspeakers are fundamentally flawed electro-mechanical devices. Our job is to discover how things go wrong, and invent solutions that ameliorate those problems as much as possible. When we finally design a product that adds zero distortion, then we will finally see a product that approaches perfection. Until then, we will continue to have an R&D division 30 persons strong.

 

Regards,

 

Patrick Butler

B&W Group North America

 

 

Thank you for the review.

I hardly ever read the subjective impressions but it was nice to learn that another magazine is performing measurements.

 

On fig.5 you can see the midrange driver's resonance peak at around 3.3KHz (green trace), right where the ear is most sensitive.

It's 5dB in amplitude and the effects on the overall frequency response are visible in both fig.1 and fig.5 (pink trace); these measurements correlate very accurately to what I have listened to at dealers and shows.

The tweeter also appears to be at least some 4dB too "hot".

 

 

As I have mentioned previously, BnW may have (probably did) make improvements both in driver and enclosure performance.

But the frequency response is far from what one would expect, at any price.

 

You've mentioned that "B&W can probably make a speaker measure anyway they want, therefore there must be a reason they have settled on a specific design"; so why don't they make it measure better?

Maybe Patrick Butler can answer that question.

 

 

As for my comments, they're just another opinion which happens not to match yours.

 

R

[Patrick Butler]

Mastering is but one part of the process. As an example, Abbey Road does recording, mixing and mastering work, or they can perform just mastering work for an artist/label. Once the performance is on "tape" and mixed, most of the sound of the performance is set, good or bad. I would certainly agree that many classical recordings are not well served by the person placing microphones or mixing.

 

Keith Johnson at Reference Recordings, or the folks at Sound Mirror in Boston (Pentatone, Harmonia Mundi, Reference Recordings) are examples of people serving the music properly with world class recordings. Sound Mirror uses our products, Keith Johnson does not. Sound Mirror, Sterling Sound, or Abbey Road use our products because they are good tools for their jobs as Recording or Mastering engineers. Studio Monitors can just as easily distort a recording as domestic loudspeakers.

 

Regards,

 

Patrick

B&W Group North America

 

 

I have quite a few recordings of "classical" music and the large majority sounds "wrong", both from a tonal balance and from an imaging perspective.

If the engineer cannot get the sound of recordings "right", I see no reason to trust their judgement when it comes to selecting speakers.

 

And, in extremis, studio monitors serve a different purpose than what's expected of speakers for domestic reproduction of recorded music: the former must present a "clear" window of the mic feed, EQ and mix, while the latter must sound "good" when reproducing recordings.

The same could be said for speakers used for testing/reviewing audio equipment.

 

R

[semente]

Hi semente,

 

Great question! I guess the question would be measure better how? Total harmonic distortion, intermodulation distortion, thermal distortion of the voice coil, distortion caused by internal reflections in the cabinet, distortion caused by diffraction, distortion caused by the resonant frequency of the basket, etc, etc etc. Frequency response is but a single measurement, and where you put the microphone will determine how closely your measurements reflect a product's actual reference response. I often shake my head at some of the measurements that we see published in magazines. As an example, there is no peak in the tweeter's response at 3.5kHz or a dip around 5.5kHz or a peak around 8kHz. Just doesn't exist in the design.

 

Is a flat response the most important goal for our engineers? Categorically not. Important? Yes- but not the most important. The aforementioned 800 Series Matrix measured flatter. However, today's 800 Series (or the last generation for that matter) are much quieter designs with overall lower levels of distortion that sound hugely improved. At the end of the day, all loudspeakers are fundamentally flawed electro-mechanical devices. Our job is to discover how things go wrong, and invent solutions that ameliorate those problems as much as possible. When we finally design a product that adds zero distortion, then we will finally see a product that approaches perfection. Until then, we will continue to have an R&D division 30 persons strong.

 

Regards,

 

Patrick Butler

B&W Group North America

 

Hi Patrick,

 

That is not much different from what I wrote previously about BnW's priorities.

I can see that a lot of this work has produced measurable as well as audible improvements; i.e. the more recent tweeter dome materials have allowed BnW to push the breakup resonance much further away from the audible range.

But it appears to me that while focusing on those causes of audible "problems" BnW is neglecting the importance of a flat response (say ±1.5dB) between 300Hz and 15KHz.

 

We all know that loudspeakers are "flawed" or, in other words, that they produce different kinds of distortion or "coloration" in orders of magnitude that are much higher than what you get from electronic equipment, which can be traced to things like cabinet, driver and port behaviour, baffle shape, as well as crossover design, etc. and that these distortions affect both the sonic "character" and performance (or "transparency") of the speaker: "tonal balance", "clarity" at low SPL, distortion at high SPL, low frequency extension, floor-bounce control and off-axis response, etc.

 

And I won't deny the importance of reducing different types of distortion, of making a speaker more "quiet" as you put it, but in my experience "tonal balance" or frequency response at the listening spot (resulting from both direct and radial dispersion) is paramount (as long as long as other distortions are kept at a reasonably low level).

"Tonal balance" can be compared to colour accuracy in an image and if you don't get it right (or flat) this will have a major impact in reproduction, or better yet in "accuracy".

 

Let's use the harpsichord as an example for it's an acoustic instruments that plays all notes with the same intensity: what do you think happens when you reproduce a good recording with a pair of speakers that do not have a flat frequency response, like this pair of Wilson Audios?

 

Frequency response, 20Hz - 20kHz (measured @ 2m, plotted @ 1m)

Top curve: on-axis response

Middle curve: 15 degrees off-axis response

Bottom curve: 30 degrees off-axis response

Listening Window, 20Hz - 20kHz (measured @ 2m, plotted @ 1m)

Response curve is an average of five measurements:

on-axis, 15 degrees left and right off-axis,

15 degrees up and down off-axis

source

 

My point is that at this asking price and expected performance level there can be no excuse or justification for a loudspeaker not to have a flat frequency response between 300Hz and 15KHz.

 

 

Moving on to subject of "independent" measurements, I chose those published by Soundstageas an example for they are expertly produced at Canada's National Research Council facilities, or to be more precise in their anechoic chamber.

 

You mention that "Frequency response is but a single measurement, and where you put the microphone will determine how closely your measurements reflect a product's actual reference response" and I agree, which is why Soundstage provides what they call a "Listening Window" averaged measurement as well as off-axis measurements (JA at Sphile's does something similar though perhaps not as accurate).

 

 

You also say that you often shake your "head at some of the measurements that we see published in magazines. As an example, there is no peak in the tweeter's response at 3.5kHz or a dip around 5.5kHz or a peak around 8kHz. Just doesn't exist in the design.".

 

I find this strange since the measurements provided by different sources are showing very similar results...

 

The first thing that comes to mind is "Why is BnW no longer publishing measurements of their speakers?"

You would be able to substantiate your claims that "independent" measurements aren't representative of "real" performance and provide the prospective buyer with valuable information.

The professional market is far more demanding and expects very descriptive specifications as well as measurements but I understand that BnW products are no longer as attractive to the professional market (your higher performing models have evolved from working tools to luxury items and are priced accordingly).

 

By the way, Soundstage Magazine reviewed the 803 Diamond in 2011.

Do you know the reason why the speaker's performance was not measured at the NRC or those measurements were not published?

 

Best,

Ric

[徐中銳]

Alright, after studying, re-studying and dissecting Ricardo's words, I'll simply quote too from Ken Ishiwata's Q&A :

How do you measure quality?

KI: Of course we have instruments. But these can only measure ‘sonic parameters’ in a static way. Instruments can only measure instantaneously – it’s like taking a still photograph of a dancer: it is precisely accurate but shows nothing of the dynamism, speed and rhythm of the dancer. Music is also dynamic. Its tone, volume, pitch and intensity continuously change. That’s why every time I work on a product I measure its quality by referencing a piece of music that I absolutely know sonically and perceptually from its original source. Only then can I relate the character of each component as part of a whole.

 

It is essential to understand what quality in original music really means. I believe this is the only way to reference the design process. You can’t just take a commercial CD as a reference point, because you can’t know the authenticity of its origins. At Marantz, we understand this importance as an absolute necessity.

Take then this photo of Sylvie Guillem :

To you, how well does she in black and white (no less) measure ?

 

Me, I've seen her Auckland performance of 2009, and would gladly see more (but she's retiring soon).

 

Similarly, I'd gladly audition B&Ws because I trust Ken's esteem of them, plus his designing of amplifiers to drive them properly. For example, his recent Bi-Amping of PM-11S3 in Floating Control configuration :

 

[semente]

Alright' date=' after studying, re-studying and dissecting [b']Ricardo[/b]'s words, I'll simply quote too from Ken Ishiwata's Q&A :

How do you measure quality?

KI: Of course we have instruments. But these can only measure ‘sonic parameters’ in a static way. Instruments can only measure instantaneously – it’s like taking a still photograph of a dancer: it is precisely accurate but shows nothing of the dynamism, speed and rhythm of the dancer. Music is also dynamic. Its tone, volume, pitch and intensity continuously change. That’s why every time I work on a product I measure its quality by referencing a piece of music that I absolutely know sonically and perceptually from its original source. Only then can I relate the character of each component as part of a whole.

 

It is essential to understand what quality in original music really means. I believe this is the only way to reference the design process. You can’t just take a commercial CD as a reference point, because you can’t know the authenticity of its origins. At Marantz, we understand this importance as an absolute necessity.

 

Let's not confuse music with reproduction of recorded music (sound).

 

Music is an art form that uses sounds as a means of expression.

It can be experienced live or reproduced in a different moment of space and time.

 

With live music one is experiencing (listening and watching) the event in real-time, we are in the presence of reality.

 

What then is the goal of music recording and reproduction system?

I would say that it is to capture as much as possible of the real event and make it's repetition as realistic as possible in a domestic environment.

 

This is how EMI once defined it in a paper called "The Pursuit of High Fidelity...":

 

Let us say, then, that High Fidelity implies the creation, in the listener's normal surroundings, of the ILLUSION of the actual performance as it would have been heard under the most favourable conditions.

 

Audio is engineering, not art; circuits designs depend on scientific theories to work.

And the purpose of measurements is to evaluate and quantify performance of certain sound parameters in a more accurate, repeatable an bias-free manner.

It isn't replace listening, but complements it.

If I am not mistaken, the measurement for Transient Intermodulation Distortion exists because the problem was detected by listening; it characterises amplifier performance in this particular parameter and helps the designer meet his goal.

 

Some designers like Daniel Weiss, Andrew Jones (ex-Pioneer/TAD) or Peter Walker (Quad) have categorically claimed that they ddon't listen to their designs but rely solely on measurements.

 

But audio is a business and to succeed (sell) manufacturers need to have some sort of "house sound" that can be identified in a shop demo.

This "house sound", design (aesthetics) and "intelligent" PR (that appeals to emotion response and confuses the reader) will produce a legion of "loyal" fans.

 

Naim, Audio Note UK and

are very efficient at this.

 

 

Take then this photo of Sylvie Guillem :

To you' date=' how well does she in [u']black and white[/u] (no less) measure ?

 

Me, I've seen her Auckland performance of 2009, and would gladly see more (but she's retiring soon).

 

 

In my view, you are asking the wrong question.

 

What I see in my screen is the reproduction of an original photograph.

The question should be : How accurate is that reproduction? (measurably)

 

Let's imagine that it was taken on Kodak TMAX 400 b&w film pushed to 1600 ASA and it's final form was a 20x30cm print in Ilford MULTIGRADE IV FB FIBER paper.

This paper print is our reality, our music.

To reproduce it with our computer displays the image will have to be scanned or digitised and then converted back to an image by our graphics card.

 

And the accuracy (and realism) of such reproduction depends on the quality of the scanner and of the display, just like the fidelity of music reproduction depends on the quality of the recording and preproduction systems.

Some aspects of this accuracy or fidelity can be measured both for imaging/video (things like dynamic range, colour accuracy, contrast, resolution, chromatic aberrations, geometric distortion, "noise) or sound.

 

The major difference is that distortions are much easier to identify in an image than in the reproduced sound of a recorded instrument or voice.

 

Here are a few examples:

 

Geometric distotion

 

 

Chromatic aberration

 

 

Corner sharpness at different aperture settings

 

 

Noise

 

Ghosting or phantom images

 

R

[徐中銳]

Eloise, rather than type you my amusing observation about AV Players today, I better save it when I'm not so busy.

 

Incidentally, a few days ago, Andrew Everard told of his « review of the Bowers & Wilkins 803 D3 speakers is out now in HiFiCritic Magazine. hificritic.com » :

 

And I should mention to readers that, because of a resolution 10 days ago, I won't ever directly respond to semente again.