DYNAUDIO C4 VS B&W 800 D3

[highfigh]

How do engineers do their designs if they believe that they can't measure relevant properties?

Such as?

If you discuss this with an EE, they're more likely to call it a filter, than a crossover. They'll be concerned with its Q, F3 or F6, the driver's impedance and yes, the phase angles. However, if you start asking about cables with properties that can't be measured and the kind with batteries that don't do anything, they're gonna look at you like you're crazy because none of the things that these cables are said to do are measurable. Engineers live in a very objective world, not the "I have a unicorn" world of high-end audio.

 

[highfigh]

True. They perform a huge number of 3D measurements in their anechoic chamber, so they know better than anyone how the speaker measures at every point in space. And it's the big picture that matters.

But that's in an anechoic chamber, which bears no resemblance to the average (or even the best) listening rooms. That's the Catch-22 of speaker testing- should it test best in a reverent space, or anechoic?

 

[Beave]

Such as?

If you discuss this with an EE, they're more likely to call it a filter, than a crossover. They'll be concerned with its Q, F3 or F6, the driver's impedance and yes, the phase angles. However, if you start asking about cables with properties that can't be measured and the kind with batteries that don't do anything, they're gonna look at you like you're crazy because none of the things that these cables are said to do are measurable. Engineers live in a very objective world, not the "I have a unicorn" world of high-end audio.

Uhm, I am an EE. Well, I was. I worked on audio electronics design and testing for a number of years. That's why I asked the question that I did.

 

[Beave]

Possibly by not slavishly adhering to the dogma surrounding the properties they can measure.

This is a total non-answer. It sounds impressive until the reader realizes it says nothing at all of substance.

 

[Beave]

WRT 'swing', think of it as a pendulum.

An inductor or capacitor will present a 90 degree shift in phase but if the value is just wrong, a second order crossover's coil that operates with minimal load below the cap's operating range presents a short. 5 Ohm to 40 Ohm jumps don't really cause problems with an amp, just output, maybe. The impedance peak at Fs keeps the driver from oscillating wildly.

We seem to be talking past each other. Perhaps because we're interpreting the other poster's claims differently?

 

[Beave]

But that's in an anechoic chamber, which bears no resemblance to the average (or even the best) listening rooms. That's the Catch-22 of speaker testing- should it test best in a reverent space, or anechoic?

Anechoic results DO bear resemblance to the sound in listening rooms in predictable ways. Toole's NRC and Harman work shows this.

Speakers should be tested anechoic, because that's calibrated and repeatable. In room tests can vary from room to room, but anechoic tests should be consistent from one chamber to another. Then the anechoic results can correlate to in-room results in typical or average rooms. Design targets for anechoic results can be derived that give consistently good in-room responses.

 

[highfigh]

Uhm, I am an EE. Well, I was. I worked on audio electronics design and testing for a number of years. That's why I asked the question that I did.

Uhm, I am an EE. Well, I was. I worked on audio electronics design and testing for a number of years. That's why I asked the question that I did.

Coulda said that before.

Your question about engineers seemed to be from someone who wasn't one.

Where did you work?

 

[Beave]

I've mentioned before that I'm a former EE, just not in this thread. I sure hope I don't have to mention it in every thread I post in!

I'd rather not say where I worked. US company that did a/v electronics, consumer and commercial use.

 

[Beave]

Such as?

If you discuss this with an EE, they're more likely to call it a filter, than a crossover. They'll be concerned with its Q, F3 or F6, the driver's impedance and yes, the phase angles. However, if you start asking about cables with properties that can't be measured and the kind with batteries that don't do anything, they're gonna look at you like you're crazy because none of the things that these cables are said to do are measurable. Engineers live in a very objective world, not the "I have a unicorn" world of high-end audio.

This actually supports why I asked the question I did. You made my point for me. He implied that they know things about sound quality that they can't measure, which is why I asked the question.

 

[<eargiant]

This is a total non-answer. It sounds impressive until the reader realizes it says nothing at all of substance.

I'm not looking to get into a back and forth. Even Gene DellaSala agrees with the line I posted. I've seen him paraphrase it in one of his articles here.

For the specific answer you're looking for you'll have to write to the folks at Steyning but I wouldn't expect a response. Why would they divulge all their R&D?

What I do know is that if Bowers & Wilkins wanted to - they could easily make their 800 series measure so perfectly that they'd make John Atkinson shudder in awe. Yet they have chosen, based on their years of research to produce the 800 Series with "tailored" (as JA says) measurements that are not exactly perfect and often scoffed at by the dogmatic adherents. Yet when you actually listen to the 800 Series, all that goes out the window.

Even with those less than perfect measurements I believe they sound about as close to the real thing as you can get. Earlier this month I made a trip to New Orleans and went to a small place that had a live Jazz band playing. The group had an upright double bass, saxophone, trumpet, trombone, piano, drums and guitar. I sat one row of tables back. As soon as they started playing my wife and I smiled because this is exactly the sound I get in my LR. These speakers perfectly convey the energy and realism of how these instruments sound in real life. It doesn't get much better than that.

At least two other members on this thread have heard this line of speakers and it seems they'd agree with me. My point is there are things that speakers can do that cannot be conveyed onto a graph (at least not yet) and there are things that can be conveyed on a graph that may not be as important as some may think. Clearly that's how the folks at Steyning see it or maybe they just agree with one of John Atkinson's Golden Rules - "All Measurements Tell Lies" so they don't get bogged down with the individual details of picture perfect measurements but rather as Art Vandelay (is that you George ) said, they look at the big picture.

As far as I'm concerned they got the big picture just right because they have one of the best speaker lines on the market. Flawed measurements and all.

Peace

 

[Beave]

So why haven't their speakers dominated Harman's blind testing?

 

[Art Vandelay]

What I do know is that if Bowers & Wilkins wanted to - they could easily make their 800 series measure so perfectly that they'd make John Atkinson shudder in awe. Yet they have chosen, based on their years of research to produce the 800 Series with "tailored" (as JA says) measurements that are not exactly perfect and often scoffed at by the dogmatic adherents. Yet when you actually listen to the 800 Series, all that goes out the window.

Take a look at the room averaged response of the 803d3 in a previous post and you'll see that there's nothing 'tailored' about it. The response is flat within a 4dB envelope. I'm sure that the engineers could design the speaker to be dead flat on-axis at 1m from the baffle, but that may or may not translate to a flat response in the listening position in a normal listening environment, which is what the engineers are trying to achieve.

FR is also just one of many parameters that make up the subjective end result. There's harmonic distortion, cumulative decay response, impulse response, phase response etc etc.

A friend of mine who owns an older B&W model also worked with the DEQX team for a while, so not surprisingly he used one of the DEQX boxes to correct the on-axis response. After a few days of listening he reverted back to native uncorrected response, which in his words sounded "far more accurate".

 

[Beave]

FR is also just one of many parameters that make up the subjective end result. There's harmonic distortion, cumulative decay response, impulse response, phase response etc etc.

Impulse response is tied to frequency response. CSD is also tied to frequency response. Phase response isn't nearly as audible as many seem to think it is. Harmonic distortion can be audible, but tying the measurements most places do to the audibility of the distortion is very difficult, if not impossible.

 

[Art Vandelay]

So why haven't their speakers dominated Harman's blind testing?

Maybe because Harmon didn't design them.

If I was to set up a pair of Revel Salon's alongside b&w 802D3's and switch between them to a blindfolded audience I reckon that most would choose the Revels. On the other hand, if I went to the trouble of getting the B&W's tow-in and rake just right, and optimised distance from room boundaries etc, the result would be very different, albeit for the select few people sitting in the sweet spot.

So what does blind testing really prove? As a EE myself I would say it can prove whatever I'm trying to prove.

 

[<eargiant]

Good points, if I recall his "tailored" comment was on the D2's which are the ones I own. If that's tailored, I'll take it any day of the week over some of the so-called "accurate" speakers I've heard. For the record, the only speakers I'd ever replace the D2's with are D3's. Why look elswhere, they get so much right that I'm hooked.

I've owned speakers that had beautifully flat FR graphs and wonderful on/off axis blah-blah-blah and while they sounded neutral they sounded nothing like real music or vocals.

 

[Beave]

Maybe because Harmon didn't design them.

If I was to set up a pair of Revel Salon's alongside b&w 802D3's and switch between them to a blindfolded audience I reckon that most would choose the Revels. On the other hand, if I went to the trouble of getting the B&W's tow-in and rake just right, and optimised distance from room boundaries etc, the result would be very different, albeit for the select few people sitting in the sweet spot.

So what does blind testing really prove? As a EE myself I would say it can prove whatever I'm trying to prove.

How do you know this without ever having performed such a test?

 

[Art Vandelay]

Impulse response is tied to frequency response. CSD is also tied to frequency response. .

F response tells us nothing about the time domain; Only freq domain.

 

[Art Vandelay]

How do you know this without ever having performed such a test?

I've participated in many such tests.

 

[Beave]

F response tells us nothing about the time domain; Only freq domain.

Not entirely true. CSD and frequency response are related in minimum phase devices such as transducers.

 

[Dennis Murphy]

Maybe because Harmon didn't design them.

If I was to set up a pair of Revel Salon's alongside b&w 802D3's and switch between them to a blindfolded audience I reckon that most would choose the Revels. On the other hand, if I went to the trouble of getting the B&W's tow-in and rake just right, and optimised distance from room boundaries etc, the result would be very different, albeit for the select few people sitting in the sweet spot.

So what does blind testing really prove? As a EE myself I would say it can prove whatever I'm trying to prove.

But you could go through the same exercise with the Revels, and their sound would improve. So I don't see where the original blindfolded comparison would be invalid. Floyd really does know what he's talking about.