Wilson Audio Duette 2 vs Sophia

[D Murphy]

I've always heard the term phase coherent to mean the phase is "correct", meaning in phase at the crossover point. Not to be at the same cycle. I've heard that described always as time coincident (or some even say phase coincident).

I don't believe the term Phase coherence is all that well defined in speakers, so there are probably many definitions.

As for waveguides, probably won't convince me of that view. I'm quite the opposite in my view. I believe waveguides are needed to produce the best sounding speakers. I've heard lots of bad speakers with waveguides, but just as many, if not more, without. My reference speakers are the Gedlee Abbey C's and I've still never heard a speaker I think is clearly better. Different at best, most are worse. It is precisely the OS waveguide that gives them their magic.

I believe the problem with the Pioneer bS22 is actually the crossover design relative to the capability of the tweeter, unrelated to the waveguide. It was crossed over too low and the tweeter distorted. When I measured it and the tower version, I was shocked to find really high distortion around the crossover region. I had thought the speaker was broken until multiple samples I tested from the line all had the problem.

Here is an example:
View attachment 32145

I wouldn't judge a waveguide by speakers like that. Listen to a JBL M22, anything from Geddes, even the Revel M126be and tell me if you still think waveguides cause coloration.

T

 

[D Murphy]

I'm happy to say we are in complete disagreement. Obviously we're not going to convince each other, but just for the record, the problem with the BS22 wasn't the tweeter, and actually, it didn't distort very much, and it wasn't crossed too low. I tried redoing the crossover with a number of approaches and crossover points, and it didn't help. It was the wave guide. That's why I replaced the tweeter with a conventional one, tricked the crossover for it, and sold a whole lot of speakers. Also, I have a Geddes Abbey sitting about 4 feet away from me. Someone sent it to me because he was running out closet space to stash speakers. If a certain other member of this forum happens to read this, he can post what he thought of it. It's actually an excellent example of the downside of wave guides. I know all of this is off topic and of limited interest, but I just can't help letting my fingers do the typing when someone posts that wave guides are a necessity. They solve some problems, and introduces others. I'm certainly not alone in this position--two of the best and most successful crossover designers I know also have problems with wave guides and the whole notion that carefully controlled dispersion is an essential part of speaker design.

 

[Matthew J Poes]

I'm happy to say we are in complete disagreement. Obviously we're not going to convince each other, but just for the record, the problem with the BS22 wasn't the tweeter, and actually, it didn't distort very much, and it wasn't crossed too low. I tried redoing the crossover with a number of approaches and crossover points, and it didn't help. It was the wave guide. That's why I replaced the tweeter with a conventional one, tricked the crossover for it, and sold a whole lot of speakers. Also, I have a Geddes Abbey sitting about 4 feet away from me. Someone sent it to me because he was running out closet space to stash speakers. If a certain other member of this forum happens to read this, he can post what he thought of it. It's actually an excellent example of the downside of wave guides. I know all of this is off topic and of limited interest, but I just can't help letting my fingers do the typing when someone posts that wave guides are a necessity. They solve some problems, and introduces others. I'm certainly not alone in this position--two of the best and most successful crossover designers I know also have problems with wave guides and the whole notion that carefully controlled dispersion is an essential part of speaker design.

Yes Dennis I think we do disagree fundamentally, but that is ok. I have the utmost respect for you, even if you dislike my favorite speakers.

I think we’ve had this pioneer discussion before. You could be right, but that then means I was sent an entire bad batch of pioneer speakers from that series. I measured the tower, bookshelf, and center. All had the same distortion rise in the crososover and it showed up on the left and right pair of each model. That measurement I posted was taken at 95dB at 2 meters. that is loud but the level was part of an output test to look at the distortion rise. Since the distortion rise as the level increases at the crossover point, I assumed it was inherent in the design and not a broken part.

I’ll certainly agree with you that those speakers were not very good sounding. I was always very disappointed in their sound. I read rave review after rave review. When I finally got my hands on them I was really underwhelmed. They just sounded lifeless.

now I’m going to pout that one of my favorite speaker designers dislikes my other favorite speaker designer’s design.

 

[Swerd]

Also, I have a Geddes Abbey sitting about 4 feet away from me. Someone sent it to me because he was running out closet space to stash speakers. If a certain other member of this forum happens to read this, he can post what he thought of it. It's actually an excellent example of the downside of wave guides.

That other forum member would be me .

I heard that speaker, first reading all about them on the GedLee web page. After reading all that, I was prepared to be happily surprised by their sound. But I wasn't. They sounded just like nearly all other speakers with horn-loaded drivers or tweeters that I've heard before – honky – megaphone-like. It wasn't subtle, I noticed it immediately and was certain I didn't like it within seconds. It didn't vary with the type of music or listening position. My only take on my reaction (and Dennis's) is that tastes in loudspeaker sound qualities can and do vary widely. I don't agree that controlled dispersion is THE answer, as opposed to designing speakers with very wide dispersion. It may work with some listeners, and with some room acoustics, but they certainly aren't for me. And I am, by far, not alone in this opinion.

The only example of a horn-type speaker I liked were Volti Vittoria speakers. I heard them once at an audio show, in a typically small somewhat crowded hotel room. Other than what Volti Audio says about them, I don't know anything about their design. They are very expensive, well beyond what I might pay. I thought they sounded effortless and were easy to listen to. There was no edginess, graininess, or honky sound that I've heard in so many other horn designs.

On the topic of time & phase correct vs. coherence, I'm quite familiar with the Vandersteen Model 3A and 2CE speakers, but not the Quattros. Most or all Vandersteens are designed to be time and phase correct. And from what I understand, they are. They can create excellent stereo images, but only if you sit in a narrow sweet spot. If you move your head horizontally, by mere inches, this effect disappears. Of course, the width of the sweet spot is frequency dependent. The mid-range frequencies, and higher, have the narrowest sweet spots. Vandersteen's use of 1st order roll-off curves contribute to this problem. Two drivers on either side of a crossover can be playing simultaneously over a wide frequency range. As soon as a listener's head moves away from a sweet spot, phase variations do occur.

For what it's worth, I've personally found that the sweet spot I hear with Dennis's designs, all of them with 4th order L/R type crossovers, create a much wider sweet spot and more stable imaging than what I've heard with Vandersteen 3A speakers. His designs do not employ standard looking 4th order crossovers where the electrical (voltage) roll-off curves are 24 dB/octave. Instead, they seek 4th order acoustic roll-off curves of the crossover network combined with the acoustic performance of the drivers. Dennis also carefully adjusts the roll-off curves on either side of a crossover to be as symmetric as possible over as wide a range as possible, usually about ±1 octave. Thus, in his designs, drivers on either side of a crossover are in phase with each other, but out of time by 1 full cycle (360° out of phase). Hence the term phase coherent as opposed to phase correct. I personally prefer this sound quality over that of any other speaker design I've known.

As usual, when it comes to discussing personal preferences in loudspeaker sound qualities… your mileage may vary.

 

[shadyJ]

That other forum member would be me .

I heard that speaker, first reading all about them on the GedLee web page. After reading all that, I was prepared to be happily surprised by their sound. But I wasn't. They sounded just like nearly all other speakers with horn-loaded drivers or tweeters that I've heard before – honky – megaphone-like. It wasn't subtle, I noticed it immediately and was certain I didn't like it within seconds. It didn't vary with the type of music or listening position. My only take on my reaction (and Dennis's) is that tastes in loudspeaker sound qualities can and do vary widely. I don't agree that controlled dispersion is THE answer, as opposed to designing speakers with very wide dispersion. It may work with some listeners, and with some room acoustics, but they certainly aren't for me. And I am, by far, not alone in this opinion.

The only example of a horn-type speaker I liked were Volti Vittoria speakers. I heard them once at an audio show, in a typically small somewhat crowded hotel room. Other than what Volti Audio says about them, I don't know anything about their design. They are very expensive, well beyond what I might pay. I thought they sounded effortless and were easy to listen to. There was no edginess, graininess, or honky sound that I've heard in so many other horn designs.

On the topic of time & phase correct vs. coherence, I'm quite familiar with the Vandersteen Model 3A and 2CE speakers, but not the Quattros. Most or all Vandersteens are designed to be time and phase correct. And from what I understand, they are. They can create excellent stereo images, but only if you sit in a narrow sweet spot. If you move your head horizontally, by mere inches, this effect disappears. Of course, the width of the sweet spot is frequency dependent. The mid-range frequencies, and higher, have the narrowest sweet spots. Vandersteen's use of 1st order roll-off curves contribute to this problem. Two drivers on either side of a crossover can be playing simultaneously over a wide frequency range. As soon as a listener's head moves away from a sweet spot, phase variations do occur.

For what it's worth, I've personally found that the sweet spot I hear with Dennis's designs, all of them with 4th order L/R type crossovers, create a much wider sweet spot and more stable imaging than what I've heard with Vandersteen 3A speakers. His designs do not employ standard looking 4th order crossovers where the electrical (voltage) roll-off curves are 24 dB/octave. Instead, they seek 4th order acoustic roll-off curves of the crossover network combined with the acoustic performance of the drivers. Dennis also carefully adjusts the roll-off curves on either side of a crossover to be as symmetric as possible over as wide a range as possible, usually about ±1 octave. Thus, in his designs, drivers on either side of a crossover are in phase with each other, but out of time by 1 full cycle (360° out of phase). Hence the term phase coherent as opposed to phase correct. I personally prefer this sound quality over that of any other speaker design I've known.

As usual, when it comes to discussing personal preferences in loudspeaker sound qualities… your mileage may vary.

Geddes intended his speakers to be listened at a hard toe-in rather than having the speakers face the listener directly, so you need to make sure you are listening with them positioned correctly.

 

[D Murphy]

Yes Dennis I think we do disagree fundamentally, but that is ok. I have the utmost respect for you, even if you dislike my favorite speakers.

I think we’ve had this pioneer discussion before. You could be right, but that then means I was sent an entire bad batch of pioneer speakers from that series. I measured the tower, bookshelf, and center. All had the same distortion rise in the crososover and it showed up on the left and right pair of each model. That measurement I posted was taken at 95dB at 2 meters. that is loud but the level was part of an output test to look at the distortion rise. Since the distortion rise as the level increases at the crossover point, I assumed it was inherent in the design and not a broken part.

I’ll certainly agree with you that those speakers were not very good sounding. I was always very disappointed in their sound. I read rave review after rave review. When I finally got my hands on them I was really underwhelmed. They just sounded lifeless.

now I’m going to pout that one of my favorite speaker designers dislikes my other favorite speaker designer’s design.

I'm glad to hear that you didn't take my late-night diatribe personally. Speaker stuff isn't worth that. Controlled dispersion is just a sore spot with me. I never measured the stock BS22 tweeter at anything like 95 dB at two meters. That's pure torture. The RAAL would also show major distortion at that level, if it survives at all. The NRC measures distortion at 90 dB at two meters, and if the speakers do very well on that tough test, a 95 dB reading is thrown in for good measure. Most magazines measure at 90 dB at only one meter. Of course, it's controversial whether you hear elevated tweeter distortion in the area of, say, 1-3%. A certain inexpensive mini-tower speaker you may be familiar with clocked in with a tweeter distortion of 12% to 18%. I haven't listened to it at deafening levels, but on normal playback I'm not sure I hear anything.

 

[Matthew J Poes]

That other forum member would be me .

I heard that speaker, first reading all about them on the GedLee web page. After reading all that, I was prepared to be happily surprised by their sound. But I wasn't. They sounded just like nearly all other speakers with horn-loaded drivers or tweeters that I've heard before – honky – megaphone-like. It wasn't subtle, I noticed it immediately and was certain I didn't like it within seconds. It didn't vary with the type of music or listening position. My only take on my reaction (and Dennis's) is that tastes in loudspeaker sound qualities can and do vary widely. I don't agree that controlled dispersion is THE answer, as opposed to designing speakers with very wide dispersion. It may work with some listeners, and with some room acoustics, but they certainly aren't for me. And I am, by far, not alone in this opinion.

The only example of a horn-type speaker I liked were Volti Vittoria speakers. I heard them once at an audio show, in a typically small somewhat crowded hotel room. Other than what Volti Audio says about them, I don't know anything about their design. They are very expensive, well beyond what I might pay. I thought they sounded effortless and were easy to listen to. There was no edginess, graininess, or honky sound that I've heard in so many other horn designs.

On the topic of time & phase correct vs. coherence, I'm quite familiar with the Vandersteen Model 3A and 2CE speakers, but not the Quattros. Most or all Vandersteens are designed to be time and phase correct. And from what I understand, they are. They can create excellent stereo images, but only if you sit in a narrow sweet spot. If you move your head horizontally, by mere inches, this effect disappears. Of course, the width of the sweet spot is frequency dependent. The mid-range frequencies, and higher, have the narrowest sweet spots. Vandersteen's use of 1st order roll-off curves contribute to this problem. Two drivers on either side of a crossover can be playing simultaneously over a wide frequency range. As soon as a listener's head moves away from a sweet spot, phase variations do occur.

For what it's worth, I've personally found that the sweet spot I hear with Dennis's designs, all of them with 4th order L/R type crossovers, create a much wider sweet spot and more stable imaging than what I've heard with Vandersteen 3A speakers. His designs do not employ standard looking 4th order crossovers where the electrical (voltage) roll-off curves are 24 dB/octave. Instead, they seek 4th order acoustic roll-off curves of the crossover network combined with the acoustic performance of the drivers. Dennis also carefully adjusts the roll-off curves on either side of a crossover to be as symmetric as possible over as wide a range as possible, usually about ±1 octave. Thus, in his designs, drivers on either side of a crossover are in phase with each other, but out of time by 1 full cycle (360° out of phase). Hence the term phase coherent as opposed to phase correct. I personally prefer this sound quality over that of any other speaker design I've known.

As usual, when it comes to discussing personal preferences in loudspeaker sound qualities… your mileage may vary.

this is very much not my experience. If the speaker had any kind of megaphone like quality I would expect that could be proven by problems in the measurements.As Dennis noted, the Geddes approach leads to beautiful measurements. While I am sure there are things we can hear that we don’t yet know how to measure, what makes a horn honk is known and defined. It was a major part of Geddes work and what he claimed to have eliminated. Which he showed through measurements and simulations.

james is right. The Geddes speakers must be listened to on the 20 degree outside listening axis. A lot of folks misunderstood what Geddes meant and toed then in 20 degrees. Doing that just puts you on axis with the speaker and into the narrow window where a cancelation dip in the waveguide causes a coloration. It disappears quickly.

I’ve had the BMR’s right next to the Abbeys and listened side by side. For some time in fact. I loved the BMR’s but I did not feel they were better than the Geddes. They were certainly more extended in the top end.

 

[Matthew J Poes]

I'm glad to hear that you didn't take my late-night diatribe personally. Speaker stuff isn't worth that. Controlled dispersion is just a sore spot with me. There was a typo in my comments about the BS22 tweeter that rendered the paragraph pretty much meaningless. I meant to say the problem with the 22 was the tweeter, not the crossover. I never measured that tweeter at anything like 95 dB at two meters. That's pure torture. The RAAL would also show major distortion at that level, if it survives at all. The NRC measures distortion at 90 dB at two meters, and if the speakers do very well on that tough test, a 95 dB reading is thrown in for good measure.

I did basically a compression sweep. 80, 85, 90, 95. I stop when the distortion rises.

I grabbed that graph because the problem was the worst/most obvious. It was high at 85 and 90, just not that high.

I think I asked you about doing a similar test on your speaker up to 90 and you asked me not to, for fear I would hurt them. I was kind to the BMR but I did measure their distortion at the equivalent of 85dB at 2 meters. I measured near field because it was in room. I remember them having very low distortion.

for me, a speakers clean maximum output is an important characteristic of its sound quality. I apparently listen pretty loud and for a long time disliked a lot of otherwise good speakers. It is likely that the biggest selling point for me on the Geddes speakers were actually its high clean output. There is no rise in distortion until about 105dB and it remains modest up to 115-120 dB Range. For me, a speaker needs to have low distortion at my 3 meter listening distance out to 105dB. Most speakers simply can’t do that. While I don’t listen at 105, that is the upper limit of the cinema standard and when I rock out, it’s in the neighborhood of the peaks.

to me, nothing ruins a good listening session or movie more than a distorting speaker.

 

[Swerd]

Geddes intended his speakers to be listened at a hard toe-in rather than having the speakers face the listener directly, so you need to make sure you are listening with them positioned correctly.

When I heard them, it was a single speaker with mono playback. I don't specifically remember what off-axis angles I positioned myself. What I typically do is walk across the speaker, about 6-8 feet away, from far off-axis on one side to far off-axis on the other side. That should have included a hard toe-in angle of 20° off-axis. I never found a spot where that horn coloration got less noticeable.

 

[Swerd]

If the speaker had any kind of megaphone like quality I would expect that could be proven by problems in the measurements. As Dennis noted, the Geddes approach leads to beautiful measurements. While I am sure there are things we can hear that we don’t yet know how to measure, what makes a horn honk is known and defined. It was a major part of Geddes work and what he claimed to have eliminated. Which he showed through measurements and simulations.

Could you post some frequency response graphs that show what makes a horn speaker honk?

 

[Matthew J Poes]

Could you post some frequency response graphs that show what makes a horn speaker honk?

Geddes was the one who made those claims, but as I understand it, his belief of what caused this "honk" was the reflections inside the waveguide. Some where what he called HOM's and others were more normal interference effects inside the horn. He once told me that HOM's are hard to display in a measurement but would basically show up as a cleaner impulse response (as these are the reflections taking place across the width of the mode over it's entire length, hence high order mode). Same for other interference effects, but they would seen as more of a repeating pattern, like comb filtering, in the measurements, as well as less hashiness (little irregularities in the response).

There is also resonances and diffraction. Diffraction is usually a bit hard to measure precisely without being able to remove it. But with diffraction based CD horns, usually one of the known effects is a strong interference effect that changes its position in the frequency plot as you move across the medial plane. If I can find a measurement of this to post, I will. I know I've seen it before and even measured it before, but I don't know if I have any on this computer. Resonances usually show up as a peak or dip and don't move with position in the far field, but change when in the nearfield (so if the resonance is caused by some shape/dimension of the waveguide and you shove the mic way inside, the resonance might be picked up differently).

I didn't suggest I have measurements to prove anything. I don't have measurements of bad vs good horns really.

 

[Matthew J Poes]

When I heard them, it was a single speaker with mono playback. I don't specifically remember what off-axis angles I positioned myself. What I typically do is walk across the speaker, about 6-8 feet away, from far off-axis on one side to far off-axis on the other side. That should have included a hard toe-in angle of 20° off-axis. I never found a spot where that horn coloration got less noticeable.

Where did the speaker come from? Just one? Do you know which version of the Abbey this was? If Dennis has this, I'd love to see measurements of it. Very likely if he could measure it's response across a few angles I could tell which version of the crossover and everything.

One thing I have always found true of the Geddes speakers (but I wouldn't consider this a coloration, def. not a horn sound) is that they are on the warm side and lack the clarity and extension of better direct radiating speakers. This is really an artifact of the DE250 and waveguides/compression drivers. Most, if not all, compression drivers do not extend out to or past 20khz, and in order for Geddes to maintain a flat DI across the frontal hemisphere, the response above 15khz was compromised a bit by shelving down. It's inherent in the waveguide itself, and to fix it, there should be a small cap on the series resistor used to raise the upper treble. The value of that (the bigger the brighter if I recall) helps bring back some of that lost upper end extension. It will never have output past 18khz though, that is the compression driver itself. The most clear difference between the BMR and the Geddes was the upper treble. Again though, for me, that wasn't a coloration.

As for the Abbey response, here is the response Geddes supplied. I haven't measured mine in years and no longer have the data. Maybe one day James will help me move them out of the theater to measure. They are pretty heavy.

and

The above is for an earlier version, but it shows you the consistency of the response at least.