+1 😉
AFAIC, the act of partaking of reproduced audio / video is predicated on the willful suspension of disbelief with which we accept the simulation of "experiencing" something out of time & space - and one only need to consider other aspects of life that are wisely disallowed as topics of conversation here to remember how easily and willingly we are fooled every day.
Hi
A coax driver is conceptually a good solution but there is still more to what goes on. I didn’t measure all of them but I did measure the most likely suspects and found the B&C 8 inch to have the smallest level of artifacts. There are two problems (at least), first is that the transition from the center pole horn to the cone body is or usually is a pretty significant discontinuity once one is well above crossover. Also, the cone works as a horn, it also has a large discontinuity at it’s edge which is also not so good “as is”.
That latter problem can be fixed by extending the horn at the same angle and line as the cone, past the cone and rounding over the last portion such as here;
SM100 | Danley Sounds Labs | Danley Sound Labs, Inc.
For the diy’r, this can be accomplished using spandex to make that shape and epoxy to make it ridged.
If the hf driver is behind the lf driver, one can also compensate the “all pass” phase and GD of a normal named type crossover above first order (where the lf section is behind the hf).
If you have a crossover program, set up say a 2nd order crossover, observe the GD, delay the hf section an amount comparable to the GD, invert it and then with a little eq find a filter shape (a non-traditional shape) where both the amplitude and phase sum flat. It takes some fiddling, but it can be done.
With the Synergy horns, there is one constant angle horn passage to deal with and the coax discontinuity is not an issue. One can use a coax driver and still get this effect as in an sm-60, shown in fig7 here;
https://www.google.com/patents/US82...X&ei=7Xm8U9GSBZGKyASl2YK4DQ&ved=0CCIQ6AEwATgK
So far as “what you hear”, these were developed using a “generation loss” recording, done on a tower up off the ground so that only the loudspeaker is captured. Surprisingly, a typical loudspeaker can only go one or two generations before sounding pretty lame, most of the Synergy horns can go 3 or 4 before reaching that point. The reasoning was that a loudspeaker should be faithful enough to the signal to be included in generation loss test and go at least a generation or two before being “bad’ sounding.
If you have a measurement microphone and 24/96 recorder, try this, it’s a much harder test than it sounds .
Also, what I have found to be a good indicator of faithful imagin is to feed the R and L speaker with the same signal (mono), ideally one hears a strong mono phantom but no R and L source is audible., something most hifi speakers do not do, they generally radiate an identity which competes with the desired image.
While we don’t sell to the hifi market, my goal is very much high fidelity but in an area where it is vastly harder to make “good sound” and stereo than in a living room and it needs to “sound the same” over a wide seating angle and span of distances . Here, the unavoidable flaws of the line array design has made this pretty easy. If you have headphones, here are a couple sound check videos of speakers larger than an sh-50 but work the same way.
This was the first public demo of the J-3, which lead to systems used in the other systems below and the video taken by one of the attendees.
Danley Sound Labs - YouTube
This system (in the scoreboard) can produce 108dB slow, A weighted at 800 feet and has a -3dB of 28Hz and is within + - 4dB in every seat.
https://www.dropbox.com/s/oyosfc3adc6j1du/20130723135350.mts
https://www.dropbox.com/s/va4mihvefqyxk24/20130723140018.mts
Another system that fits in a scoreboard
https://www.dropbox.com/s/tnsw5mb4v5vdlwq/20120726122124.mts
Spartan Stadium's New Sound System - YouTube
A less powerful distributed system (the things along the roof line) on a very cold (5 F) windy day;
https://www.dropbox.com/s/1nhx980w24ehxls/20131208154446.mts
A demo that lead to a system being installed now, this is just three cabinets (the little black glob under the scoreboard) replacing a very large concert sound system from a large mfr.
Penn State Demo.MOV - YouTube
We don’t advertise but by this fall, we will have the speaker systems at half of the country’s 100,000+ seat stadiums and many smaller ones. The coolest part is, because the sources in line arrays radiate independently (and don’t add coherently but radiate an interference pattern) they require MANY more drivers, amplifiers, dsp and boxes to do the same job (very good for mfrs, but bad for the customer and listeners) .
Line arrays have much less directivity in comparison, a smaller usable working distance, sound different everywhere and smear time horribly (one arrival from each source) and are very prone to wind problems because they radiate an interference pattern produced by the same source spacing issues I was referring to. In the living room, these things can exist too on a smaller scale.
Best,
Tom
A coax driver is conceptually a good solution but there is still more to what goes on. I didn’t measure all of them but I did measure the most likely suspects and found the B&C 8 inch to have the smallest level of artifacts. There are two problems (at least), first is that the transition from the center pole horn to the cone body is or usually is a pretty significant discontinuity once one is well above crossover. Also, the cone works as a horn, it also has a large discontinuity at it’s edge which is also not so good “as is”.
That latter problem can be fixed by extending the horn at the same angle and line as the cone, past the cone and rounding over the last portion such as here;
SM100 | Danley Sounds Labs | Danley Sound Labs, Inc.
For the diy’r, this can be accomplished using spandex to make that shape and epoxy to make it ridged.
If the hf driver is behind the lf driver, one can also compensate the “all pass” phase and GD of a normal named type crossover above first order (where the lf section is behind the hf).
If you have a crossover program, set up say a 2nd order crossover, observe the GD, delay the hf section an amount comparable to the GD, invert it and then with a little eq find a filter shape (a non-traditional shape) where both the amplitude and phase sum flat. It takes some fiddling, but it can be done.
With the Synergy horns, there is one constant angle horn passage to deal with and the coax discontinuity is not an issue. One can use a coax driver and still get this effect as in an sm-60, shown in fig7 here;
https://www.google.com/patents/US82...X&ei=7Xm8U9GSBZGKyASl2YK4DQ&ved=0CCIQ6AEwATgK
So far as “what you hear”, these were developed using a “generation loss” recording, done on a tower up off the ground so that only the loudspeaker is captured. Surprisingly, a typical loudspeaker can only go one or two generations before sounding pretty lame, most of the Synergy horns can go 3 or 4 before reaching that point. The reasoning was that a loudspeaker should be faithful enough to the signal to be included in generation loss test and go at least a generation or two before being “bad’ sounding.
If you have a measurement microphone and 24/96 recorder, try this, it’s a much harder test than it sounds .
Also, what I have found to be a good indicator of faithful imagin is to feed the R and L speaker with the same signal (mono), ideally one hears a strong mono phantom but no R and L source is audible., something most hifi speakers do not do, they generally radiate an identity which competes with the desired image.
While we don’t sell to the hifi market, my goal is very much high fidelity but in an area where it is vastly harder to make “good sound” and stereo than in a living room and it needs to “sound the same” over a wide seating angle and span of distances . Here, the unavoidable flaws of the line array design has made this pretty easy. If you have headphones, here are a couple sound check videos of speakers larger than an sh-50 but work the same way.
This was the first public demo of the J-3, which lead to systems used in the other systems below and the video taken by one of the attendees.
Danley Sound Labs - YouTube
This system (in the scoreboard) can produce 108dB slow, A weighted at 800 feet and has a -3dB of 28Hz and is within + - 4dB in every seat.
https://www.dropbox.com/s/oyosfc3adc6j1du/20130723135350.mts
https://www.dropbox.com/s/va4mihvefqyxk24/20130723140018.mts
Another system that fits in a scoreboard
https://www.dropbox.com/s/tnsw5mb4v5vdlwq/20120726122124.mts
Spartan Stadium's New Sound System - YouTube
A less powerful distributed system (the things along the roof line) on a very cold (5 F) windy day;
https://www.dropbox.com/s/1nhx980w24ehxls/20131208154446.mts
A demo that lead to a system being installed now, this is just three cabinets (the little black glob under the scoreboard) replacing a very large concert sound system from a large mfr.
Penn State Demo.MOV - YouTube
We don’t advertise but by this fall, we will have the speaker systems at half of the country’s 100,000+ seat stadiums and many smaller ones. The coolest part is, because the sources in line arrays radiate independently (and don’t add coherently but radiate an interference pattern) they require MANY more drivers, amplifiers, dsp and boxes to do the same job (very good for mfrs, but bad for the customer and listeners) .
Line arrays have much less directivity in comparison, a smaller usable working distance, sound different everywhere and smear time horribly (one arrival from each source) and are very prone to wind problems because they radiate an interference pattern produced by the same source spacing issues I was referring to. In the living room, these things can exist too on a smaller scale.
Best,
Tom
interesting "technical paper" from Genelec (coaxials and DSP): http://www.genelec.com/documents/other/Genelec 8260A Technical Paper.pdf
Nope, graaf! You've just fallen into the usual "Best" trap! This is why HiFi goes round in circles. 😀
"Different" is a better word to use. Or "What is good about a Tannoy and its siblings, and what is worse?"
I like them too, as it goes, but let's not think you've reached the end of the road. It's an exponential horn at heart, with all the efficiency gains that go with them. But you can do horns in lots of ways. The Tractrix, aka the pseudosphere, for instance has kinder edge effects. Which brings us back to Tom Danley's solution in a way:
An externally hosted image should be here but it was not working when we last tested it.
All variations on a sphere. It's a spatial filter which has a spatial fourier transform just like a linear crossover filter. Every curve has different spatial dispersion and frequency response and group delay.
Here's an interesting idea which actually comes out of the stable solutions in string theory. Within a spherical geometry lurks a toroid or doughnut shape. This is hard to grasp, but within a, say, spherical power response lurks a toroidal group delay. The sphere and the toroid compliment each other.
An externally hosted image should be here but it was not working when we last tested it.
A Black Hole might seem a bit off-the-wall, but it's all about the geometry. 😎
Attachments
Last edited:
? just asking
oh, I see the misunderstanding 🙂 "best!" in my post was just an abbreviation of "best regards!" 🙂
ok, what is exactly worse then? About Tannoy, or about contemporary KEF's UniQ (which differs significantly both from Tannoy and the original UniQ), or about Genelec's MDC?
please tell me 🙂
Last edited:
I'm not suggesting any solutions at all at this stage, just assembling the pieces that might be relevant. That's what you do to be creative. You then sleep on it. 😀
Coincident drivers solve the lobing problem with loudspeakers. But they also cause a greater interaction between the drivers for one thing. They also beam. And you haven't taken away the crossover:
I used to think you couldn't hear the group delay in that 4th order slope, which BTW isn't the only slope or frequency response you can use. But you can! We can hear the difference between closed box LR2 rolloff and reflex LR4 rolloff, for instance. It's time delay!
Here's how nature does the maths of dispersion from a spherical aperture:
And here's what happens if you confine an impulse in a square box, which could be a physical box, or a filter with sharp edges. It becomes periodic or rings.
No conclusions at this stage. 🙂
Coincident drivers solve the lobing problem with loudspeakers. But they also cause a greater interaction between the drivers for one thing. They also beam. And you haven't taken away the crossover:
An externally hosted image should be here but it was not working when we last tested it.
I used to think you couldn't hear the group delay in that 4th order slope, which BTW isn't the only slope or frequency response you can use. But you can! We can hear the difference between closed box LR2 rolloff and reflex LR4 rolloff, for instance. It's time delay!
Here's how nature does the maths of dispersion from a spherical aperture:
An externally hosted image should be here but it was not working when we last tested it.
And here's what happens if you confine an impulse in a square box, which could be a physical box, or a filter with sharp edges. It becomes periodic or rings.
An externally hosted image should be here but it was not working when we last tested it.
No conclusions at this stage. 🙂
What do You mean? Does it make them worse in any way? In what way?
So You think that somewhat higher but constant directivity is a problem, right? Do You prefer omnidirectional designs?
NO, you're looking for conclusions already, grasshopper! 
Your mind has to be very quiet to do this. We are trying to look at the right things here.
Here's that ol' Allison IC20 speaker again, designed to mount against the wall. You need two of these cabinets. Left and right, as usual. It's very clever at many levels. Because it's designed to interact with the room very well. The filter is likely to be third order butterworth if Roy Allison was as good as I think he was. Which is a constant power design. It's also a cylindrical source, rather than a spherical one, so power falls off more slowly with distance.
But being just two speakers, the best we can hope for is the stereo illusion. You'd need many more drivers to get all the elements to recreate a total directional sound field. That becomes more of a holographic problem.
BTW, you will notice it is another solution to the lobing problem. Perhaps better than the D'Appolito MTM. And it has elements of holography, because the apparent sound source is not actually coincident with the individual drivers.
Your mind has to be very quiet to do this. We are trying to look at the right things here.
Here's that ol' Allison IC20 speaker again, designed to mount against the wall. You need two of these cabinets. Left and right, as usual. It's very clever at many levels. Because it's designed to interact with the room very well. The filter is likely to be third order butterworth if Roy Allison was as good as I think he was. Which is a constant power design. It's also a cylindrical source, rather than a spherical one, so power falls off more slowly with distance.
But being just two speakers, the best we can hope for is the stereo illusion. You'd need many more drivers to get all the elements to recreate a total directional sound field. That becomes more of a holographic problem.
BTW, you will notice it is another solution to the lobing problem. Perhaps better than the D'Appolito MTM. And it has elements of holography, because the apparent sound source is not actually coincident with the individual drivers.
Attachments
Last edited:
NO, you're looking for conclusions already, grasshopper!
well, from here it looks like You are making statements I am just asking 🙂
that's all
so You say "room interaction", Allison, "holography", so perhaps You might be interested in reading this: http://www.diyaudio.com/forums/multi-way/121385-advantages-floor-coupled-up-firing-speakers.html
the original title of the thread was "Loudspeakers and room as a system"
I think you'll find that the Tannoy is a conical horn at heart.
Hence its constant directivity characteristic requiring a 6dB/oct boost from 5kHz or so up.
I don't care what sort of horn it is really. It's a horn for sure though. Which is an acoustic impedance matching device that makes for greater efficiency in coupling the diaphragm to the air. Different shapes have different characteristics on frequency response, spatial dispersion and group delay. And a flat baffle is a sort of horn too, to a certain limit of parameters, if you follow?
graaf, putting a bass close to a boundary like the floor makes it couple better to the room. You can hear it. That's not hard, but an element, for sure. In fact the Allison effect is named after the man who described it all accurately. 🙂
Here's another one. Gaussian response works better than most things in filters and driver response. And apertures. It's because the far field diffraction pattern is also Gaussian if you do the spatial transform.
Again Roy Allison built this into his drivers. There's a lot of nice things in his designs. 😎
graaf, putting a bass close to a boundary like the floor makes it couple better to the room. You can hear it. That's not hard, but an element, for sure. In fact the Allison effect is named after the man who described it all accurately. 🙂
Here's another one. Gaussian response works better than most things in filters and driver response. And apertures. It's because the far field diffraction pattern is also Gaussian if you do the spatial transform.
Again Roy Allison built this into his drivers. There's a lot of nice things in his designs. 😎
Attachments
Last edited:
ok, so what are the disadvantages of a coaxial design like the latest generation UniQ or Genelec's MDC?
Hi graaf
I didn’t measure Tannoy or KEF drivers, I don’t recall that they were available as a raw driver but that was 9 years ago when I designed the SH-100 (now an SM-100 and it’s brothers) and a lot of water has gone under the bridge since then. I did measure all of the coax drivers I could get from the larger driver mfr’s and some others since then.
I know that we did a generation loss recording on at least one Tannoy speaker as well as most of our competitors but I don’t remember what model it was, that was also a long time ago now.
It was kind of funny as the recordings were much more a reality check for us than anything else as we were going in a different direction where there were no “road signs”.
The Genelec paper you linked points out some of the things I had found back then about the discontinuity’s, that stuff is real.
Best,
Tom
Unfortunately they are still unavailable as raw drivers but a determined DIYer can buy UniQs, only in the form of a custom installation kit (the CI series) and then just throw away all the cheap plastic they come with.
They are much more expensive to buy that way of course, and I wonder whether such an option is worth the extra money.
Why would flat power response be a goal, except in the most padded of cells?Replace that with smooth power response as indicated by the literature, and it's possible to get most of the rest of it with careful design.
I've played with basically every concentric design on the market except for the Genelec MDC and Danley's genius riff on the concept*: several generations of Uni-Q and Dual Concentric. TAD/Pioneer CST; the pro-audio ones cobbled together from parts bins by B&C, Radian, Beyma, etc. "Better" is indeed the right word.
The thing the current Uni-Q and Dual (both designed under the same lead engineer, Mark Dodd) get right are the mouth transition. The CST units are almost as good in that area, because the tweeter is actually a cone and the edge flows neatly into the surround. The others aren't in the same ballpark, and sound like it.
As for driver interaction, that of course happens any time multiple drivers are on a flat baffle as well. Limit the excursion of the concentric's cone driver with careful crossover design, and it's just not an issue.
You're getting too bogged down in the little itty bitty details here IMO. 🙂
We are after seeing the big picture for ALL speakers! 😎
Here's a very interesting tweeter, which has some gentle horn loading. I don't think the stepped taper matters one bit. Something smooth would work as well.
H1499-06 27TBCD/GB-DXT
Here's a tidy woofer too:
H1659-08 U22REX/P-SL
Now I don't much care whether that is engineered as concentric or a two way. Because I'm looking at other things I like in the response.
We are after seeing the big picture for ALL speakers! 😎
Here's a very interesting tweeter, which has some gentle horn loading. I don't think the stepped taper matters one bit. Something smooth would work as well.
H1499-06 27TBCD/GB-DXT
Here's a tidy woofer too:
H1659-08 U22REX/P-SL
Now I don't much care whether that is engineered as concentric or a two way. Because I'm looking at other things I like in the response.
Sadly The KEF Coax efforts are erm Inadequate DC designs.
'Built to a price point you know' 🙄 Too many insurmountable issues with it's sound, by my experiences. A dead end product.
As digression; there is a bespoke purpose designed/tailored Active crossover for Tannoy Alnico Drivers.. It works well, effectively a full ranger result.
Have one / recommended. Pure DIY though.
DSP 'may' be used for similar result, dunno, haven't tried.
Soooo much talk in Audio and sparse substance.
Clearly hot air is the Energy of the Audio arena.
Thousands of Speaker maker 'Businesses' have come and gone since the mid 70's
Yet it continues.. Astounding IMO.
Indeed, commercial HiFi is a very cynical business. I was once visiting a PA company. I had already been taken aback how the young folks in the dingy solder bay were breathing huge amounts of lead fumes, making crossovers.
I was then listening to their new active two way speaker designed for use in pubs. I thought it sounded rough, and wondered if it needed better drivers. The response surprised me: "Ah, it doesn't matter. We get those from Poland for a few quid. And we can arrange a good review in a magazine!"
I really don't know how much difference it makes to lose ceramic ferrites in coils and speaker magnets. Quite a lot I expect. The hysteresis curve is quite alarming. AlNiCo largely died away when world spot prices of Cobalt went through the roof. And Neodymium magnets corrode rather badly, don't they?
I have a 8" paper Sony driver that is remarkably like a Tannoy driver. It sounds really good too. Nothing exotic, but the minimal dustcap and cloth surround works really well.
I was then listening to their new active two way speaker designed for use in pubs. I thought it sounded rough, and wondered if it needed better drivers. The response surprised me: "Ah, it doesn't matter. We get those from Poland for a few quid. And we can arrange a good review in a magazine!"
I really don't know how much difference it makes to lose ceramic ferrites in coils and speaker magnets. Quite a lot I expect. The hysteresis curve is quite alarming. AlNiCo largely died away when world spot prices of Cobalt went through the roof. And Neodymium magnets corrode rather badly, don't they?
I have a 8" paper Sony driver that is remarkably like a Tannoy driver. It sounds really good too. Nothing exotic, but the minimal dustcap and cloth surround works really well.
Attachments
- Status
- Not open for further replies.