[Rocket]

Hi Dave,

Glad you don't mind my comments.

The phl drivers sound very revealing but do seem to have a bit dry sound and i know a couple of audiophiles in perth who don't like the sound they produce.  My current speakers use a focal 7k4211db mid/bass driver and although they aren't as transparent as the phl's they do sound a bit warmer imo.

I'm basically going to keep my current speakers just in case i don't like the phl tower i'm building.

It's a pity i don't live in the states as i'd like to hear your 1801 speakers.  

Regards

Rod

[Rocket]

Deleted due to double post.

[skrivis]

Most excellent.  This is progress in the right direction.  You are saving me a whole bunch of work at the keyboard, and this should read very linear.  I will dispense with further questions, and provide my 2 cents on the issue.  My basic overview is thus:

"You can't have yer' cake and eat it too".

Yep, the only way the well damped material can convert the sound energy into heat is movement.  Insulation fibers are the easiest to understand in this regard.  The do flex when encountering sound waves, and ...

Perhaps small amounts of movement within the cone material are not as bad as larger movements of cone or surround with a cone material that doesn't have much internal damping?

A very good actual example of this is exists in the response curve for the W18E driver on the SEAS page. The surround material thickness and type determines the response slightly above 2500hz. The relative dip between 2500hz and cone resonance at 4200hz is the surround material moving out of phase with the driver. However, this also means that the CONE of the W18 driver remains a piston up to this frequency range. The stiff edge of the cone at high acceleration pushes the surround into "whiplash".

The surround must be considered part of the radiating surface...

Is a "perfect" cone that requires the surround to "whiplash" really what we want?

[skrivis]

If you want to use a 1st order acoustic crossover with a 7" driver at 2500hz, that cone needs to have flexible material, in a flexible cone to absorb the sound energy. It won't have resonance.

Which goes along well with my preference for 1st-order crossovers and crossover points that are well away from the midrange.

[skrivis]

Better drivers... there we get into my opinion that "stiff enough" drivers with good damping characteristics are better than stiffer drivers that don't have self-damping.

What truly stiff drivers have you actually measured/tested/implemented/used?  What's your a/b comparison?

I've heard a couple of systems using the SEAS Excel drivers and they actually sounded pretty good.

For the rest, I'm simply speaking from my knowledge of how drivers behave, based on what I've read over the years.

I've also used the same speaker cabinets for about 15 years, much of the time with the same crossover, but different drivers. The drivers were all quite similar except for the cone material. There were a couple of iterations of polypropylene, then kevlar composite, and finally paper. (I guess we could call this "lignin composite" hehe)

I've made general observations about the sound of these cone materials over the years. (From theory as well as listening.) My observations jibe quite closely with Lynn Olson's, so I feel they're valid to some extent.

I've also read quite a bit of Ted Jordan's work, and I have to say he probably knows more about drivers than anyone else. I highly recommend his work to anyone interested in this subject.

He arrives at a slightly different conclusion than I do. Jordan feels that the proper surround can control any misbehavior by the cone. That's ok for the upper frequencies, but I'm less convinced about this for bass. Jordan seems fairly ok with conventional cones for bass though...

I've heard the Jordan modules, and they sounded quite good, although they seem rather fragile. If you accept everything that Jordan says, then his drivers are the best solution.

However, I've also heard Hiquphon tweeters a lot, and they take an opposite approach, one of using well-damped material. I would say that the Hiquphon tweeters sound better than the Jordan drivers, and they're more robust too.

So I'm not speaking as someone who has designed drivers or anything. These are just my opinions. Other people seem to share my viewpoint.

In fact, if I were trying to think of a way of improving a "paper" cone, I'd want to make it a bit stiffer, but not sacrifice the internal damping. How would we do that? Well, I think the ScanSpeak slit-paper cones are an excellent idea.  A lot of people I've talked to regard these drivers as the best on the planet. Hmmm....

[skrivis]

John,

Thanks for the commentary but it's uncool to post a commercial for your speakers in another manufacturer's circle.  In fairness, Dave should edit your last paragraph.  I'm not big on rules but this one seems to make sense.  

Cheers,

Jack

I think I may have actually mentioned the Xd's earlier.

They're an excellent example of the stiff cone, steep crossover paradigm that David mentioned.

David also said he had some interest in DEQX and wanted to learn more.

I don't really see the Xd's as competition for David's designs. If nothing else, people who appreciate the beauty of David's cabinetry will think the Xd's look butt-ugly.

They're also more expensive, and they don't let you use whichever amplifiers you like best, plus they use Class D amps and I have some real reservations about those.

I have always preferred to deal with small companies where I stand a chance of getting personal service; sometimes from the owner or engineer. That was a nice thing about Fried Products, and it's one of the nice things about Audio by Van Alstine. It's a nice thing about Ellis Audio. Look at the conversation I've been having with David (and others too). We've been talking about things in great detail, and David has provided very good insight into his mind as a designer. Now try doing the same with NHT...

John may offer very good service, but he's an NHT dealer, and it won't be the same as dealing directly with NHT. NHT simply can't provide that "small company service" that I'm talking about either - they're too big.

Finally, I was a drum roadie. I've got several friends who are drummers. I've done work in the studio and heard what a drum kit sounds like there.

The Hiquphon tweeters do a *superlative* job with cymbals. They just sound right. The metal dome tweeters I've heard over the years do not sound this good.

Given the digital equipment used in the Xd's, and especially the Class D amps, I find it quite possible that the Xd's put out plenty of trash that will excite any resonances in the metal dome tweeters.

I'm willing to listen to the Xd's, but I'm skeptical.

So that's my rebuttal to John.

I do agree that he has a tendency to pop up and tout the Xd's, but maybe I'm more forgiving than some...

[skrivis]

When measured, the Accuton C95 nasty response bubble around 4.2khz is darn near on par with the SEAS W18 drivers.  However, there is something else happening here.  I really don't know what the issue IS, but when my CLIO jig is working, it will be explored further.

I am not surprised at the very small quantity of Accuton drivers in the hifi marketplace.  They are expensive - very expensive.  Given the quantity of production for hifi speakers and the necessary profit margins, using Accuton is a very quest ...

The Accuton drivers aren't too bad. They're more expensive than Hiquphon or some of the Vifa domes, and I think they're about the same as the ScanSpeak Revelator. The one that's really pricey os the ScanSpeak R2904/7000.

Your comments about the lack of interest in live music are valid. Most people do not have that as a reference.

Whether live music really _should_ be the reference might be arguable. I think it does depend upon the type of music and performance.

I can say that my current stereo system does a good job with giving me the same "character" as what I've heard in the recording studio.

[John Ashman]

Less rebutting, more listening

I have a customer who has Hiquwhatevers in his DEQX project upon which he's working, so I'll have to go listen to them, I've never heard them before. Maybe I'll drag over them thar Xds  

I did hear the B&W diamond tweeter and it was interesting.  I didn't quite like it, in my short listen, as much as the Xd tweeter right next to it, but that is probably an implementational thing, diamond *should* be a nice step forward.  Unless you hear two drivers in an "all other things more or less equal" situation, it's hard to tell where the driver ends and the implementation begins.  

Maybe I just like the *sound* of metal tweeters more than the transparency, but so far..............

[skrivis]

With a cone, the surround is in the right place to control the bad behavior. With a dome, there's nothing there to control it...

I believe this comment hits the target very close to the center. I believe the difference between many good tweeters (fabric or metal) is the surround. The very valuable dampening of a surround on cone drivers is never overlooked - especially on very well damped (i.e. soft) cones. As the frequency rises, such cones outer edge will begin to flop in the opposite direction of the inner cone movement. The well damped surround is obviously needed to control the loose floppy motion of the outer edge of the driver cones.

I thought maybe I should discuss what I meant a bit more.

The surround on a dome tweeter can _never_ control just the diaphragm's behavior. It will effect the entire structure, including the voice coil. I think it's purpose is akin to that of the spider on a cone driver.

The problem areas in a cone or dome tend to be the areas farthest from the voice coil.

With a cone the farthest area is attached to the surround, and the surround plays a vital role in controlling the cone.

With a dome, the farthest point on the diaphragm from the voice coil is the center of the dome. There is no surround there to damp anything. The center of the dome basically does what it wants to if there is breakup.

That's why I feel that materials such as doped silk work well. There's enough damping within the material itself to avoid problems with the dome center. It's possible that metal domes with a coating of damping material gain enough damping that they avoid major problems, but the figures I've seen seem to indicate that a metal diaphragm with a coating of damping material is heavier than a flexible material with damping. More mass lowers the resonant frequency, by increasing inertia. More mass is not a good thing for a tweeter, and not always a good thing for a woofer either.

My understanding of surrounds on dome tweeters is that they provide centering, and also a seal. They may offer some control or damping too, but it's really control or damping of the voice coil - not the dome itself.

[John Ashman]

Stuart, the problem is that the resonances in a soft dome will be much higher in amplitude within the useful bandwidth of the speaker.  A rigid structure vibrates less and the resonant energy is disipated as a spike outside of the audible range.  But with a soft dome, it will be oscillating within the bandwidth.  So, a soft dome will lose some detail and will also add resonances, though they aren't all that offensive, not like the oilcan resonance would be *if* it were in the audible bandwidth.  

This is why I alway bitch about kevlar.  If you look it on an interferometry scan, the fabric is undulating badly compared to a ridid material and its doing it where you'd hear it.  

Soft domes are still in use mainly because some are bothered by or think they are bothered by metal domes.  NHT, for instance used to use both, soft domes for the low end models, but have switched to all metal and won't ever go back.  Revel, who actually *does* interferometry resonance scans of their drivers uses rigid domes.  In fact, the reality is that companies that use hard domes are generally seeking to meet objective performance goals, not cater to audiophile perceptions.  Maybe that's good, maybe that's bad.  But there are a lot of objectivist audio companies that measure, measure, meausre, set goals and keep trying for perfection.  PSB, Paradigm, (actually, that whole Canadian group) NHT, Revel, Thiel, etc, etc.  They all use metal domes and probably will until stuff like diamond is available cheaply.  Too be honest, I think much of the "fatigue" people experience is because they see a metal tweeter and expect to be fatigued.  Most are so good now, I'd be surprised if most people could tell with a blindfold except that the metals will generally be more realistic on cymbals and other HF content.

[skrivis]

Stuart, the problem is that the resonances in a soft dome will be much higher in amplitude within the useful bandwidth of the speaker.  A rigid structure vibrates less and the resonant energy is disipated as a spike outside of the audible range.  But with a soft dome, it will be oscillating within the bandwidth.  So, a soft dome will lose some detail and will also add resonances, though they aren't all that offensive, not like the oilcan resonance would be *if* it were in the audible bandwidth.  

This is ...

How are the resonances from a damped soft dome greater in magnitude than those with a metal dome? A damped system tends to have a low Q, while things like metal domes will have a very high Q at resonance.

Please explain "the resonant energy is disipated as a spike outside of the audible range."

So you think that a metal dome can put out tons of trash at 30KHz and it won't have any effect because it's beyond the audible range? Ever heard of intermodulation?

It's especially a concern with the Xd's because I don't know the quality of the DEQX digital implementation and how much hi-freq trash it puts out. The Class D amps will likely put out a bunch of noise too, and excite any resonances in the metal dome tweeters.

Floyd O'Toole/NRC-inspired designers? Well.... if that's your cup of tea, that's fine.

[John Ashman]

How are the resonances from a damped soft dome greater in magnitude than those with a metal dome? A damped system tends to have a low Q, while things like metal domes will have a very high Q at resonance.

Defined "damped".  You mean "lossy"?  A fabric dome tweeter isn't really "damped" as I define the word. A cone that is sprayed with a resin or something that is designed to inhibit resonance is "damped".  How well it is "damped" depends on how it is done.  Fabric is not rigid, therefore it vibrates and undulates easily.  Rather than being a high Q resonant spike, the undulations cause wider bandwidth resonances, aka distortion IN bandwidth.  Yes, you could argue that a 10dB spike in energy could be problematic except that you can't hear it.  And each year of your life puts it further away from audibility.  

Please explain "the resonant energy is disipated as a spike outside of the audible range."

IOW, the only real tendency to resonate in a metal driver is at actual one resonance peak.  It has very low loss.  Sure, there might be some small resonances in the audible bandwidth, but they will be very low compared to those of a fabric design.   With fabric, however, you have flexibility combined with tensile strength, the same thing that plagues kevlar.  It *will* resonate, but across a wide bandwidth and it will be in the audible range mostly, unlike a metal tweeter.  

So you think that a metal dome can put out tons of trash at 30KHz and it won't have any effect because it's beyond the audible range? Ever heard of intermodulation?

So you think a soft dome resonating throughout the audible range won't have any effect on it?  The fabric isn't dissipating all that much as heat, it's dissipating it as sound.  Fabrics will have higher distortion figures than metal drivers.  I don't hear the peak, at least in better metal tweeters.  I've heard some tweeters with some 10kHz peaks, such as those from B&W, but that's motor, not the diaphragm.  The Diamond tweeter has the same problem.  

It's especially a concern with the Xd's because I don't know the quality of the DEQX digital implementation and how much hi-freq trash it puts out. The Class D amps will likely put out a bunch of noise too, and excite any resonances in the metal dome tweeters.

A lot of people who haven't heard Xd have this "concern".  Those who have heard it do not.  At this point, you're wildly accusing new technology based on your perceptions, not knowledge of the design involved.  I can tell you this, actively amping the tweeter sure beats running through a bunch of resistors and capacitors.  

Floyd O'Toole/NRC-inspired designers? Well.... if that's your cup of tea, that's fine.

Yep, pretty much.  Not always, but at least they haven't lost site of the ideal goal of a speaker - transparency.  Most "high-end" speaker companies are just screwing around like Baskin-Robbins does with ice cream.  Thank goodness too, or I'd be forced to listen to listen to SETs with horns and that would be horrifying.

[skrivis]

Defined "damped". You mean "lossy"? A fabric dome tweeter isn't really "damped" as I define the word. A cone that is sprayed with a resin or something that is designed to inhibit resonance is "damped". How well it is "damped" depends on how it is done. Fabric is not rigid, therefore it vibrates and undulates easily. Rather than being a high Q resonant spike, the undulations cause wider bandwidth resonances, aka distortion IN bandwidth. Yes, you could argue that a 10dB spike in energy could be problematic except that you can't hear it. And each year of your life puts it further away from audibility.

In the case of real diaphragm materials, yes, damped equals lossy.

I get the impression that you think soft dome tweeters are just made out of fabric. Perhaps some are, but the better ones use a doped or coated fabric. (Doped silk is common with tweeters from ScanSpeak, Vifa, Dynaudio, Morel, and Hiquphon)

The end result is a dome that is rigid yet has good internal damping. It doesn't go into uncontrolled high Q breakup like a metal dome will.

We don't yet have the perfect material. So metal is very stiff, but not perfectly stiff. It also exhbits very poor internal damping. (This is why they make bells out of metal instead of coated fabric.)

Metal diaphragms _will_ exhibit breakup, since we're not dealing with perfectly stiff materials. When they do, there isn't any internal damping to control the resonance.

You'll see the center of the dome start to misbehave because it's farthest from the control of the voice coil. Since metal isn't internally damped, the breakup will be very well propagated to the edge (the voice coil), then reflected back into the dome, etc.

If there's a 10 dB spike due to dome breakup, you'll be able to hear the results, whether the spike itself is in the audible range or not. Just because it isn't directly audible does not mean you won't hear its effect.

n"

Please explain "the resonant energy is disipated as a spike outside of the audible range."

IOW, the only real tendency to resonate in a metal driver is at actual one resonance peak. It has very low loss. Sure, there might be some small resonances in the audible bandwidth, but they will be very low compared to those of a fabric design. With fabric, however, you have flexibility combined with tensile strength, the same thing that plagues kevlar. It *will* resonate, but across a wide bandwidth and it will be in the audible range mostly, unlike a metal tweeter.

Metal dome breakup will be exhibited at whatever point the stiffness of the material is inadequate to compensate for the mass of the dome. Since the metal has very little internal damping, it's going to be pretty nasty when it lets go. (The exact same thing can be said about any material. The point involved will differ, and the amount of internal damping will influence the magnitude of the breakup or resonance.)

Kevlar composite cones have some limitations, but I'm not seeing how flexibility combined with tensile strength are responsible.

So if I make a cone out of something flexible with little tensile strength I won't have any problems? Or do I want something stiff with no tensile strength?

 

you think that a metal dome can put out tons of trash at 30KHz and it won't have any effect because it's beyond the audible range? Ever heard of intermodulation?

So you think a soft dome resonating throughout the audible range won't have any effect on it? The fabric isn't dissipating all that much as heat, it's dissipating it as sound. Fabrics will have higher distortion figures than metal drivers. I don't hear the peak, at least in better metal tweeters. I've

You didn't answer my question, but we'll let that pass.

I'm not convinced that a soft dome is going to resonate or breakup throughout the audible range. Furthermore, any such resonance will be of lower Q and likely less obnoxious.

Low amounts of distortion across a wide band are likely to cause less problems than lots of distortion in a few places, yet quoted measurements will often represent an average figure...

A lot of people who haven't heard Xd have this "concern". Those who have heard it do not. At this point, you're wildly accusing new technology based on your perceptions, not knowledge of the design involved. I can tell you this, actively amping the tweeter sure beats running through a bunch of resistors and capacitors.

This could well be. However, Class D amps work by switching at a high frequency and then putting a low pass filter at the output in the hopes keeping it out of your speakers. I was simply pointing out that there is a real chance of the amp exciting any resonances in the tweeters.

Yep, pretty much.  Not always, but at least they haven't lost site of the ideal goal of a speaker - transparency.  Most "high-end" speaker companies are just screwing around like Baskin-Robbins does with ice cream.  Thank goodness too, or I'd be forced to listen to listen to SETs with horns and that would be horrifying.

How do you measure transparency?

And you're saying that Baskin-Robbins screws around with ice cream?

[John Ashman]

In the case of real diaphragm materials, yes, damped equals lossy.

Which is why lossy materials lose resolution and aren't really sufficiently pistonic.  Lossy isn't actually a good thing - it's two wrongs attempting to make a right.  

The end result is a dome that is rigid yet has good internal damping. It doesn't go into uncontrolled high Q breakup like a metal dome will.

But won't have the clarity.  Internal damping absorbs music too.  

We don't yet have the perfect material. So metal is very stiff, but not perfectly stiff. It also exhbits very poor internal damping. (This is why they make bells out of metal instead of coated fabric.)

Note that they don't make bridges out of fabric either.  Poor rigidity means lots of resonances, low Q, perhaps, but lots of it everywhere, not just one spike.  It doesn't need to be well damped in places it's not really resonating.  Note that bells produce one primary note.  With an aluminum dome, that fundamental is at 20+KHz with the harmonics being above that even.  

Metal diaphragms _will_ exhibit breakup, since we're not dealing with perfectly stiff materials. When they do, there isn't any internal damping to control the resonance.

It doesn't matter if it's not audible.  You say it is, but have you ever just considered that you like a lossy, softer version of reality?  

You'll see the center of the dome start to misbehave because it's farthest from the control of the voice coil. Since metal isn't internally damped, the breakup will be very well propagated to the edge (the voice coil), then reflected back into the dome, etc.

Yes, but metal misbehaves dramatically less in the audible range.  

Just because it isn't directly audible does not mean you won't hear its effect.

Maybe then you'll need to define "audible" again, because if it's not directly audible, it sounds more like an imagined event.  

Kevlar composite cones have some limitations, but I'm not seeing how flexibility combined with tensile strength are responsible.

Simple - tensile strength means poor internal damping in one direction.  Therefore, that forces any energy to be radiated instead of absorbed.  Self-damping materials require a bit of flexibility, as the material stretches and compresses, it gives off energy as heat.  High tensile strength means poor damping and therefore, it forces the kevlar to undulate, something it can easily do since it's not rigid.  With a rigid material, that energy is released as a high Q peak out of the useful bandwidth rather than *within* the useful bandwidth.  

So if I make a cone out of something flexible with little tensile strength I won't have any problems? Or do I want something stiff with no tensile strength?

An ideal cone would be so stiff that an resonance is very high in the bandwidth, and you can't easily do that with a lossy, self-damped material like paper or poly or cloth.  Paper and poly are both flexible *and* have good damping (relatively low tensile strength compared to metals), so they don't "ring" and are well behaved *but* lose resolution.  I am not aware of a material that is stiff, but with no tensile strength.  I'm not sure that's possible exactly.  I'm trying to think of an example, but it's not coming to me.  

You didn't answer my question, but we'll let that pass.

My point is that metal domes resonate almost exclusively in the inaudible range whereas soft domes resonate mostly in the audible range.  But you're acting as one is forgivable and the other not.  I'd rather not *hear* the distortion.  It will exist, but who cares if you can't hear it?

I'm not convinced that a soft dome is going to resonate or breakup throughout the audible range. Furthermore, any such resonance will be of lower Q and likely less obnoxious.

Think of this then.  Metal is extremely rigid.  this forces the resonance point to be much higher in frequency, barely to beyond 20kHz.  Soft domes are *not* rigid, therefore, will resonate at a lower frequency, ergo, below 20kHz.  That is down in the audible range.  If you think metal deforms in the center of a dome, wait til you see what a soft dome does.

Low amounts of distortion across a wide band are likely to cause less problems than lots of distortion in a few places, yet quoted measurements will often represent an average figure...

Again, doesn't matter if its inaudible or isn't excited in the first place.  

How do you measure transparency?

Transparency equals measured accuracy + low inband distortion + time/phase accuracy + coherent dispersion.  All of that is measureable.  They almost need to have a "tranparency index" that weighs and combines these factors, except that there'd be no industry wide agreement on how to do that.

[skrivis]

Transparency equals measured accuracy + low inband distortion + time/phase accuracy + coherent dispersion.  All of that is measureable.  They almost need to have a "tranparency index" that weighs and combines these factors, except that there'd be no industry wide agreement on how to do that.

The O'Toole/NRC-related designers don't seem to adhere to those criteria in that they don't seem to worry about time/phase accuracy.

Low inband distortion doesn't always tell the whole story either. Out of band distortion can be a problem too. Low overall distortion is a nice ideal, but there may be other things that are more important. The way distortion is measured may also mask some nasty behaviors...

A big speaker system with a ton of drivers for each frequency range may have vanishingly low measured distortion, but utterly incapable of properly passing a transient.

A system could show excellent distortion figures from 20Hz-20KHz, and frighten away bats and small pets because the tweeter puts out lots of trash above 20KHz.

Metal diaphragms _will_ exhibit breakup, since we're not dealing with perfectly stiff materials. When they do, there isn't any internal damping to control the resonance.

It doesn't matter if it's not audible. You say it is, but have you ever just considered that you like a lossy, softer version of reality?

My "lossy, softer version of reality" if it actually is so, matches very well with what I hear in the studio.

If your amplifier oscillates at 30KHz, you won't hear it, but your tweeter will, and it will likely burn out. That's perhaps a worst case of the audibility of hi freq trash.

I really suggest that you look into what effects signal above the audible range has. Otherwise you're just kind of burying your head in the sand and ignoring it.

As for the whole stiff vs. less stiff thing, my contention all along is that some materials are "stiff enough" to get the job done while having good internal damping so that when they do get into trouble it isn't the end of the world.

The figures that I've seen show that metal diaphragms with damping applied mass more than composite materials where the damping is part of the structure. I'm thinking mainly of doped silk domes and paper or wood pulp cones. The added mass will reduce resolution. (I'm sure it isn't always the case that metal plus damping weighs more, but it's something I'm going to look at more closely.)

I think I can also say that we're not necessarily talking about night and day differences here. There's room for both viewpoints. I'm also quite willing to listen to the Xd's and admit I'm wrong if they do as well as you say. It's one of the things on my to-do list, and I would have already listened to them if there were a local dealer.

Getting back more on topic, I do feel that the Ellis Audio 1801's are a good design.  I may not totally agree with David, but I do feel his design has merit. My preference for his style of doing business means that I would be more comfortable with a product from him than with one from a much larger company.

"You pays your money and makes your choices."

[John Ashman]

The O'Toole/NRC-related designers don't seem to adhere to those criteria in that they don't seem to worry about time/phase accuracy.

Of course not, because it's the hardest to achieve and has the least amount of gain when you do.  And it has the worst side effects - higher distortion, more lobing, more complex crossover, limited driver choices, etc, etc, etc.  We're far more senstive to the latter problems than time/phase, that really needs to be the last thing to be done and digital makes it possible without nasty side effects.  

Low inband distortion doesn't always tell the whole story either. Out of band distortion can be a problem too. Low overall distortion is a nice ideal, but there may be other things that are more important. The way distortion is measured may also mask some nasty behaviors...

Maybe.  Sometimes it's where the distortion lies more than how much, but a really good speaker is going to be as low distortion as possible.  Digital and rigid drivers also makes that more possible.  

A big speaker system with a ton of drivers for each frequency range may have vanishingly low measured distortion, but utterly incapable of properly passing a transient.

Actually, they generally do transients better, but with more lobing and other issues.  Steep digital crossovers prevent the need for lots of drivers to achieve low distortion/high output (except for the bass - you still need capacity there.)

A system could show excellent distortion figures from 20Hz-20KHz, and frighten away bats and small pets because the tweeter puts out lots of trash above 20KHz.

Still doesn't matter if you can't hear it!  

My "lossy, softer version of reality" if it actually is so, matches very well with what I hear in the studio.

Maybe.  Our local studio has some Genelecs that are SO soft, they make Al Di Meola's Kiss My Axe sound like Muzak.  No sense of dynamics, dull, soft, warm.  I bet every CD coming out of there is harsh, bright, aggressive.  But real life has a lot of "snap" to it with percussive transients.  I had a guy tell me that my NHTs made a horn sound "brassy".  "AND?!?!?"  

If your amplifier oscillates at 30KHz, you won't hear it, but your tweeter will, and it will likely burn out. That's perhaps a worst case of the audibility of hi freq trash.

Sure, and if your car runs out of oil, the engine will fail.  That same amp would blow a soft dome too.  SACD has an *enormous* amount of high frequency "trash" and yet audiophiles love the sound, saying it's more "analog" and they want more extended tweeters so they can "hear" all the nasty effects of the noise shaping.  God bless their little audiophile hearts!  

I really suggest that you look into what effects signal above the audible range has. Otherwise you're just kind of burying your head in the sand and ignoring it.

Not really.  If a tweeter sounds better to me, it's better for me.  I can listen to the NHT's relatively cheap metal tweeter all day and yet some expensive metal and soft dome tweeters annoy me.  Certainly cheap soft domes annoy me.  No detal *and* a lot of distortion.  It's the quality of the tweeter and the implementation that counts.  The hiqophons might sound great, but if they can do cymbals well, they'd be the first soft dome I've heard that can.  

As for the whole stiff vs. less stiff thing, my contention all along is that some materials are "stiff enough" to get the job done while having good internal damping so that when they do get into trouble it isn't the end of the world.

I understand that, but what you seemed to saying is that metal has an issue with resonances at the center of the dome and all I was pointing out is that soft domes are, by nature, far more problematic so you can't really use this as a negative for metal.  

The figures that I've seen show that metal diaphragms with damping applied mass more than composite materials where the damping is part of the structure. I'm thinking mainly of doped silk domes and paper or wood pulp cones. The added mass will reduce resolution. (I'm sure it isn't always the case that metal plus damping weighs more, but it's something I'm going to look at more closely.)

Mass doesn't reduce resolution so much as frequency response.  The future lies in either stuff like diamond, beryllium or very fine vapor deposited laminated structures where multiple layers equal the thickness of a single metal dome, but are more rigid with the resonance far higher in the frequency range.  If you really dampen a soft dome, you're going to make it too heavy and it may not resonate, but it will roll off.  

I think I can also say that we're not necessarily talking about night and day differences here. There's room for both viewpoints. I'm also quite willing to listen to the Xd's and admit I'm wrong if they do as well as you say. It's one of the things on my to-do list, and I would have already listened to them if there were a local dealer.

Agreed. I like the sound of many soft domes, but prefer what I believe to be the added realism of a good metal dome.  I don't discount that there are soft domes more to my taste, but most of the speaker brands I like happen to use metal.  I am picking up Triad, so I'll get to hear some nice ScanSpeaks though.  I just find that there is a lot of distrust, almost anger towards metal domes.  It's like women, you can have a bunch of bad experiences, but if you take that baggage forward, you'll end up missing out (when a nice metal-equipped speaker crosses your path).  

Getting back more on topic, I do feel that the Ellis Audio 1801's are a good design. I may not totally agree with David, but I do feel his design has merit. My preference for his style of doing business means that I would be more comfortable with a product from him than with one from a much larger company.

Sure it's a good design, basically.  The only thing is, for you, the W18 has a big 4kHz spike, something like 15-20dB as I recall, which I'm sure is notched rather than entirely avoided.  You will get some of that in the sound, it's unavoidable, but the resolution below that will be fantastic.  It will likely have a big of a brightness to it, possibly even a little more fatiguing than if it were poly or paper.  But that's the usual price for low midrange distortion.  I remember hearing the Veldyne 661s "the lowest distortion speaker in the world" and I couldn't listen to a full song, not even a light jazz song (I'm sure Tool would have sent them into overload) without leaving.  But with the right notch and a relatively low and steep crossover, I'm sure it would be fine.  NHT feels that the SEAS cones are unusable in passive analog situations.  I'm not sure I agree with it, but I understand why they'd say it.  The use a 110dB/octave at 2000 Hz to avoid ringing.  With a 24dB/octave crossover, it would have to be down below 1000Hz, even more like 500Hz, which is why a notch filter must be in there.  Of course, NHT doesn't like using high parts count crossovers either, so it's more of a difference of design theory, which is fine.  I would be a little surprised, though, to see you not liking metal domes because they have a 20kHz+ spike, but then being okay with a metal midrange that has a 4kHz spike.  There is a paper or poly version of that driver, I believe, that would almost drop in, but without the notch filter.  You'd lose some resolution, but possibly get more of the sound you like.

[skrivis]

The O'Toole/NRC-related designers don't seem to adhere to those criteria in that they don't seem to worry about time/phase accuracy.

Of course not, because it's the hardest to achieve and has the least amount of gain when you do. And it has the worst side effects - higher distortion, more lobing, more complex crossover, limited driver choices, etc, etc, etc. We're far more senstive to the latter problems than time/phase, that really needs to be the last thing to be done and digital makes it possible without nasty side effects.

I'm solidly on the side of time/phase accuracy being very important. It's one of the reasons I'm interested in DEQX.

Lobing is only an issue if the user doesn't know about it or doesn't know where the lobes are. There's a definite sweet spot with most speakers and that's where I sit to listen.

There is some talk about power response being supremely important because it determines what the room-reflected sound will be like and that sound is a major component of what you hear. I'm not so sure that you need to base your design on a desire for flat frequency and power response. It seems like the wrong way to start in my opinion.

More complex crossover? Not always. My current speakers have 1 cap, one coil, and 2 resistors in each crossover. I'm getting a pair of speakers soon that have the identical crossover and drivers in a much larger cabinet with a longer transmission line.

More limited choice of drivers? Yep. Why is that bad? Good drivers are available, so you use those.

I tend to feel that time/phase accuracy needs to be the first thing, rather than the last. As Pat McGinty says," Our speakers are designed in the time domain FIRST and, as it turns out, when a speaker's timing is correct, the frequency domain just plain lays down and falls into place."

Sure it's a good design, basically.  The only thing is, for you, the W18 has a big 4kHz spike, something like 15-20dB as I recall, which I'm sure is notched rather than entirely avoided.  You will get some of that in the sound, it's unavoidable, but the resolution below that will be fantastic.  It will likely have a big of a brightness to it, possibly even a little more fatiguing than if it were poly or paper.  But that's the usual price for low midrange distortion.  I remember hearing the Veldyne 661s "the ...

It's possible that the W18 would not be acceptable for me. I was just trying to point out that I respect David's design skills, and, as a designer and businessman, I have a lot of confidence in him.

There are a number of things about NHT's design that I like, but there are a few areas where I'm skeptical. The proof is in the pudding, so that will have to wait for an audition of them. Other people have voiced similar concerns, about the Xd's and the technology they use, so I'm not alone.

Also, you have to admit that David's cabinets are far prettier than the Xd's.

[John Ashman]

Lobing is only an issue if the user doesn't know about it or doesn't know where the lobes are. There's a definite sweet spot with most speakers and that's where I sit to listen.

Well, you also have to be the right height for your speakers sometimes too!   It also tends to affect "power response" which may or may not be a negative, but I'm pretty sure it's not a positive

There is some talk about power response being supremely important because it determines what the room-reflected sound will be like and that sound is a major component of what you hear. I'm not so sure that you need to base your design on a desire for flat frequency and power response. It seems like the wrong way to start in my opinion.

It's a balancing act, though digital has clear advantages.  I used to think that it would just cause problems having wide dispersion, but when you think about it, speakers are inherently wide dispersion, just not at all frequencies, so "wide dispersion" generally means "coherent dispersion" which is to say that all the frequencies will be a part of the reflections, not just some of them.  This could lead to some brightness, but it is easier to deaden this with some room treatment or a rug or something.  

More complex crossover? Not always. My current speakers have 1 cap, one coil, and 2 resistors in each crossover.

Generally, yes, 1st order designs are very complex to do *right* because every driver peak needs to be addressed as do the rolloffs so you don't have wide peaks and dips in the overall response.  That's if you want accuracy as well, anyway.  

More limited choice of drivers? Yep. Why is that bad? Good drivers are available, so you use those.  

Some of the best drivers, such as Accutons or SEAS magesiums aren't appropriate for 1st order designs.  They need steep crossovers to sound their best.  

I tend to feel that time/phase accuracy needs to be the first thing, rather than the last. As Pat McGinty says," Our speakers are designed in the time domain FIRST and, as it turns out, when a speaker's timing is correct, the frequency domain just plain lays down and falls into place."

That's fine, except that you *will* get an increase in motor/cone distortions, lots of lobing, difficulty in maintaining accuracy, lowered SPL etc, etc.  You really need something like DEQX to do phase/time accuracy properly.  I don't like the sound of any of the 1st order speakers I've ever heard, except maybe Thiel 3.6s which are done very well.  I really dislike Vandersteens.  Meadowlarks were "okay", but nothing about which to get excited.  

There are a number of things about NHT's design that I like, but there are a few areas where I'm skeptical. The proof is in the pudding, so that will have to wait for an audition of them. Other people have voiced similar concerns, about the Xd's and the technology they use, so I'm not alone.

Well, it's good to be skeptical.  It's like someone advertising an anti-gravity car.  I'd be skeptical too.  But it does work because DEQX works.  They could have made it even better, but it would have doubled or tripled the price to get a modest improvement.  As it is, the acoustic design is spectacular, especially with the rigid cabinets, low diffraction baffles, choice of drivers, etc.  

Also, you have to admit that David's cabinets are far prettier than the Xd's.

I'll give you that they're more "traditionally" pretty  However, I happen to think Xd is gorgeous in a more Italian "Ducati" way.  I'm not totally thrilled with the color combo, but I think the styling is tremendous and at least differentiated.  I guess it depends on the trappings of your home and your taste.

[skrivis]

I'll give you that they're more "traditionally" pretty  However, I happen to think Xd is gorgeous in a more Italian "Ducati" way.  I'm not totally thrilled with the color combo, but I think the styling is tremendous and at least differentiated.  I guess it depends on the trappings of your home and your taste.

My father was a part-time cabinetmaker, so I got used to fine woodwork.

As for the time/phase coherent 1st-order crossover issue, I don't think you need a complex crossover to get good results. You do need to carefully design the system, and you do need to choose your drivers wisely, but you need to do that with any speaker system. Linearity, dynamics, and power handling need not be an issue either - with the proper drivers.

DEQX or a similar system offers a number of advantages, if it's done correctly. However, it's more complex and more expensive. There could  be a tendency to skimp on drivers and "correct" them so the end result measures well and the final price of the system is acceptable.

Multi-amping, which is a requirement of DEQX, can also have advantages. But it does require multiple channels of amplification, and there again there could be a tendency to skimp on each channel.

There is also a school of thought that says that 1st-order series crossovers produce a beneficial interaction between drivers.

As I said, we'll see what happens when I finally get to sit in front of a pair of Xd's. I also hope you do get a chance to hear the Hiquphon tweeters. I think you'll be pleasantly surprised.

[John Ashman]

DEQX or a similar system offers a number of advantages, if it's done correctly. However, it's more complex and more expensive. There could be a tendency to skimp on drivers and "correct" them so the end result measures well and the final price of the system is acceptable.

Trust me, if you skimp on the drivers, it will be obvious.  I did a modified Thiel system with Morel tweeters and an NHT woofer (both much better than the original Thiel) and it sounded fantastic.  Xd killed it though.  And the systems I did with cheaper drivers were nothing about which to get excited.

Multi-amping, which is a requirement of DEQX, can also have advantages. But it does require multiple channels of amplification, and there again there could be a tendency to skimp on each channel.

Modern amps are good enough that "skimping" is a matter of personal belief.  Even if you put in a decent 6-channel amp (or digital amps) instead of 6 expensive monoblocks it's going to be 19 steps forward instead of 20.  

There is also a school of thought that says that 1st-order series crossovers produce a beneficial interaction between drivers.

That school of thought is that each driver has a "sound" and lack a coherent dispersion and that you want to blend those sounds so that it is "seamless".  However, with a rigid pistonic driver, you don't really have a sound at all.  Or very little.  Xd is doing .2% midrange distortion at ~97dB.  It has almost no sound.  Therefore, people can't tell where the tweeter ends and the midrange begins, even with a 110dB/octave crossover.  Same with the sub.  All I hear all day long is the word "seamless" over and over.  IOW, the better and more rigid the driver and the steeper the crossover and the more you use the FR below the beaming point, the less need there is for blending of any type.   Done right, drivers really don't have a "sound", they become a nearly ideal transducer.