Except that all this is contrary to physics and measurements that I have made. Turbulence in a CD was studied (at B&C) and it was found that reducing the turbulence by an order of magnitude (smoothing corners etc.) had no measureable effect on the performance.
Turbulence in an oscillating flow is, by the way, almost nonexistent. That's because turbulence takes time to establish and with the flow constantly changing it cannot be established. It has been shown to exist to some degree in woofer ports (a good reason for not using them) but this is a very slow flow change when compared to a CD.
Heat loss in air is simply not ever a factor and even if it were, it would be linear, so no "distortion" would be created. I have looked at the thermal aspects of drivers from every direction and can find no real effects except the obvious ones of long term heating of the voice coil from the current and the change in the Re that results. This can be shown to have a huge effect on a passive crossover - non-existent for an active one.
I am curious . What do you measure which cannot be duplicated by a resistor and a capacitor?
Unfortunately you are right. But don't apply "statistics" here. I mean a thousand random ears is not better than one good ear. You only need to find one ear that can hear a difference then you have to conclude that it is audible, even if a thousand others can't.
And for things like this, you don't "listen" to the sound, but you "feel" the sound. If the listener doesn't understand what I'm talking about, there's no way he can "hear" the difference.
Have you read my previous post, where I stated about "small differences" and "correlation"? What do you think about it?
For example, what if I can hear sound difference between A and B which happen to have one ppm thd difference. Do you require that I should be able to hear one ppm? No, that's wrong. May be something is wrong in the measurement, or in the understanding of how things work. What important is whether this is consistent or not, have positive correlation or not.
My finding is, I can hear obvious differences which when measured, the difference is too small to be considered audible. But this should not disqualify if consistency is maintained.
Earl is right, that when DBLT has null result, it is not important anymore whether there is difference or not, it can be ignored. I agree a hundred percent, except that I want my ears to be the decision maker, not those ears they used to come to a conclusion that tube and solid state sounded the same.
My experience is that it is all audible under certain conditions, usually not with complex music, but rather with individual sounds such as human voice or other similar sounds .No doubt about that [except for magnet difference, which I have no experience with, in a controlled environment]. What I doubt is if they are immeasurable. Are they?
I'm pretty sure they are all measurable [phase, non-linear distortion, IMD], when they are audible. But I'm also sure that the measured difference will be small. But as long as there is CORRELATION between measured performance and perceived performance, then the subjective quantity doesn't matter [for those who can hear it of course]. For those who can't hear it, they will require that you can hear the distortion difference. That's a terrible "Math" imo.
My gut feeling is that boldname is correct. Not from my understanding of Physics or CD construction, but only from listening experience.
Consider the above as a phenomenon, and boldname is trying to explain the reason for the phenomenon existence.
You really need to sum together all the losses. I have only simplistically stated the physical phenomena going on. It is not contrary to the physics. If you have data, demonstrate that on an equal playing field that a CD is better at low level detail than a comparable dome tweeter, from what is available.
You cannot really argue away the problem of potentially greater low level information multiple losses from a CD unless you engineer to more than compensate for the losses. You can start to compensate by using huge magnets on the CD and many other things. In many applications it does not matter.
My belief is the best dome available will exceed any CD available on published impulse testing results. And any overshoot and ringing is a loss, ie distortion. Put a good Be dome 1" tweeter up against any CD. If you can find an available 1" CD that matches the current SEAS and Scanspeak domes then I would be interested.
Both CD and direct drivers have their place and both can be used on wave guides and horns. Each wins in its own arena, which may be live music repro, PA reinforcement and home hifi, and it can be neck and neck i.e no outright winner across all applications. Dome drivers withoutthe phase plug are far easier to design and to maximise their SQ performance.
I still like both options and will carry on reviewing the CD possibilities as the new wave of drivers especially with new diaphragm/dome materials roll out
Last edited:
I don't understand why horn/WG should be excluded. In my effort to explain the phenomenon, horn/WG is relevant...
For me, when a signal is distorted, fine details are lost. For example, when you go down to an underground tunnel, you may hear the sound of a water drop. Some people might say "wow, that's a detailed sound of water drop". But really? The sound was amplified, but with distortion. If you put your ears close to a water drop, you will know what detail is.
Now, can a horn/WG or CD mechanism be free from distortion? I don't think so. It is again [as always] about small difference which some people considered as negligible...
Taking tweeter WG as example, I found that straight WG has less tendency to ruin the sound than a curved WG. It is actually always better in term of micro sound quality to have no WG, but the sound is just too weak.
I assume that when you add the WG, you REDIRECT the off-axis wave into on-axis, and this "change" something. That's why currently I'm using felt or foam to match dispersion/directivity with assumption that I don't redirect the sound, but I absorb/delay the off-axis waves.
In the electrical domain you cannot attenuate reflections and diffraction, but you can acoustically.
I am sorry, but I think that the burden of proof is on the claim. I am just telling you what I know from physics and experience. Take it or leave it.
PS. You guys are using "distortion" in an extremely wide context that includes linear distortion. Linear distortion can be corrected and always is in a CD (no CD works without EQ.) So it is usual to reserve the term "distortion" for only the nonlinear stuff. If you don't understand this comment then you shouldn't be arguing about the physics.
Last edited:
Sorry to throw my statistician hat on, but, even then you couldn't reject the null that XYZ was inaudible because for any given test, run on 1000 people, you should and likely would observe some "by chance" that got it right. That in and of itself wouldn't prove they could actually hear XYZ [consistently] and it wasn't just random despite their own belief in their hearing. They would have to be able to repeatedly prove their hearing superiority.
I think debates about DBLTs are pointless as both sides of the camp are thoroughly entrenched. I am only sharing my view on the stats. See Nasim Taleb's Fooled by Randomness.
Last edited:
Yes, after you found that there are some people who "by chance" got it right, you have to remove the others and concentrate on these new subset. Do it until there is no more "by chance" or "random" or "belief". That's the right way to understand/implement statistics.
I was trying to give a hint. I stated to focus on phenomenon, not the explanation about the cause of the phenomenon. If this concept is understood, one can possibly find out that what really happened [the phenomenon] is that the implementations have always been wrong, i.e. introducing too much of linear distortion. So if better implementation is carried out, such phenomenon wouldn't exist.
But then, comes the next argument: "Show us the good implementation"
There is much more mechanisms beside Le(X) what current drive can negate. I have never measured a driver that has not had benefit from current drive, including woofers that have three shorting rings and very low voice coil's resting point inductance. I have never seen a driver with flat Le(X) either.
http://www.current-drive.info/6 :
Yes both sides are entrenched, but only one side is entirely justified by the science. It's just like global warming.
Taleb's books (all of them) are indeed outstanding but if you have to explain what you did above then they would be well over the posters head.
Le(x) is not the major factor at HFs since the x is so small, but the current Le(i) is not. Once one reduces the major factor the rest is not necessarily significant whether you can measure it or not. Just because you can measure something does not mean its audible. This is especially true of nonlinearities.
Current drive is not new. There are usually good reasons why something doesn't survive the test of time. Horns/waveguides have, current drive hasn't.
Last edited:
Once the fine detail is lost agreed it is lost. and the direct dome inverted or otherwise will be at its best from close up. A small spherical wave guide can marvelously improve the projection without I believe losing much detail. Adding foam or felt has to be judicious as it absorbs some wanted HF material against unwanted generated HOMS when used in a WG or horn. And that is another area of debate. Geddes has done a great job minimising any short fall of performance. We all respect and recognise that. But one does not need to damp down any HOMS in a straight dome tweeter, but then you do not get the projection and directivity and more powerful output from increased sensitivity. Short wave guides or horns have less of a problem with HOMS.
The reason for leaving wave guides and horns out of this posting as they complicated the comparison, It was purely for the debate started on the principle of CD and direct drivers. We were looking merely at direct versus indirect drivers.
I have found round wave guides can give some of both worlds of WG and horns perhaps with less coloration, while computed WG profiles and horns can give you constant directivity, or a whole range of radiation patterns or other useful changes to the sound projection. This while increasing sensitivity. This has become prevalent with the Genelec Behringer and other offerings so waveguides are now of course quite common place.
I like both wave guide and horn options, and I would use a CD if I wanted high output level or efficency for small DHT and SS amps but probably stick with domes for treble upper mid using WG or Horn for powerful amps.
Lynns excellent project is directed at these DHT amps that really demand a CD for the mid to mid treble. Geddes line is similarly esteemed and can put out with big amps.
Earl, thanks for the info on the experiments at B&C. The experiments themselves as well as the results are interesting, although I will say I’m not surprised by them. For compression drivers, at least those that work from the low mids on up, we simply can’t think of oscillating-air velocity in the same sense as that of a linear fluid flow (a vector quantity) where laminar or turbulent flow could be supported. You’re absolutely right in that turbulence cannot occur in an oscillating ‘flow’.
Besides, the space between the diaphragm and phase plug as well as the space inside the annular slits of the plug, is so small compared to the wavelengths being propagated, that if indeed turbulence could occur it would only affect frequencies way up in the ultrasonic range at over 100KHz.