I think this is similar to the Toole tests where he finds that significant speaker resonances are audible and show up in frequency response at every angle you measure. Hence power response is a good way to look for resonances. At the same time the dips at crossover points or the particular directivity of each driver don't seem to be a big factor in the listening tests.
Flat power response is seldom achieved without rising on-axis response, so really flat power response can have a negative correlation to the rankings. (Flat power response tends to sound bad.)
A lot of the well ranked systems in Toole/Olive tests are 3 ways with power response holes at the crossover points along with a downhill power trend.
So power response smoothness is important while power response flatness is not.
David S.
Flat power response is seldom achieved without rising on-axis response, so really flat power response can have a negative correlation to the rankings. (Flat power response tends to sound bad.)
A lot of the well ranked systems in Toole/Olive tests are 3 ways with power response holes at the crossover points along with a downhill power trend.
So power response smoothness is important while power response flatness is not.
David S.
Pity it's 1/3rd (?) Octave bands... It might give an idea of the overall tonal balance, but no idea whatsoever about whether the response is smooth or ragged.
IMHO narrow band smoothness is more important and a better indicator of perceived quality than overall tonal balance, especially at higher frequencies. Overall tonal balance error also lends itself to fixing with EQ while narrowband raggedness typically doesn't, unless it's it's just one or two discrete resonances that can be accurately compensated on an individual basis.
Last edited:
Hello,
here's a frequency measurement of Line Magnetic 555 replicas, post 9!
Line Magnetic triple 5 - Verstärker, Lautsprecher, Zubehör - Analog-Forum
All these measurements giving the same substantiated imagination of sound characteristics like a brainwave measurement regarding to intelligence.
This thread is invigorating and fertile like a cultural exchange between vegetarians and cannibals about food passions! Space enough for a billion posts.
Many music lovers getting their ideas/visions of hearing-receiving-'reliving' music realized with such old stuff...after a long (and expensive paid) journey for gaining experience with modern HiFi - equipment. Modern Technology gave us - underpinned and proveable - 'better-' stuff. May be, but if there's a yardstick for sheer musical involvment, the old WE horns are conveying more than everything else. But perceiving this is first of all a question of own personality and mentality. Quite a few people having heavy weather with it, because it's far beyond the usual(often boring) criterions of HiFi - qualities... technicans and theoreticians particularly..
here's a frequency measurement of Line Magnetic 555 replicas, post 9!
Line Magnetic triple 5 - Verstärker, Lautsprecher, Zubehör - Analog-Forum
All these measurements giving the same substantiated imagination of sound characteristics like a brainwave measurement regarding to intelligence.
This thread is invigorating and fertile like a cultural exchange between vegetarians and cannibals about food passions! Space enough for a billion posts.
Many music lovers getting their ideas/visions of hearing-receiving-'reliving' music realized with such old stuff...after a long (and expensive paid) journey for gaining experience with modern HiFi - equipment. Modern Technology gave us - underpinned and proveable - 'better-' stuff. May be, but if there's a yardstick for sheer musical involvment, the old WE horns are conveying more than everything else. But perceiving this is first of all a question of own personality and mentality. Quite a few people having heavy weather with it, because it's far beyond the usual(often boring) criterions of HiFi - qualities... technicans and theoreticians particularly..
Sorry, should have used '....' not quotes.
The exact statement is:
I was not trying to be incendiary.
That is speculation. Perhaps true, perhaps not. One of the big problems I can tell you, is that the people you refer to and the equipment are not usually in the same place, and when they are, they are there to do specific work - at least this is my arms length understanding of it.
You don't need to measure, test and get data when someone tells you that a lion is about to attack you. The lion is coming anyhow...
_-_-bear
Thanks!
I am reposting the curves. These were done on plane-wave tubes. The bottom curves compare 7V and 9V through the field coil.
The low-end wiggles may indicate that the PWT was under-damped at those frequencies.
I wish there was an impedance sweep, too...
Attachments
Last edited:
expecting anything less from that substantial chunk of metal(s) is ..... not logical
whenever it's made
whenever it's made
Thanks for reposing that Bill. It does look great, I just wonder what other, later drivers would look like under the same conditions?
I'd love to see something like Art's driver test done on a series of old and new drivers. That would give a real level playing field for comparison.
I'd love to see something like Art's driver test done on a series of old and new drivers. That would give a real level playing field for comparison.
Agreed. You could probably make better drivers than the WE ones and get them to perform similarly on the same horn. The real problem is combining such a large horn with something on top. That midrange is gorgeous though and probably where the magic comes from.
The range I heard them in best was about 200-1000Hz. That's the sweet spot. And a very important range, for sure.
I would love to see/hear some side-by-side comparisons. But they would all have to be 1 3/8 threaded drivers, or you would have to redesign the horn flare. Adapters would mess up the carefully controlled expo expansion.
IMO, that range would be no sweat for a long list of easily obtainable threaded drivers with phenolic 'phragms. Who's got a 15a horn/clone? Let's try this out!
Hi
I wouldn’t think you would be able to find a modern driver that equals the WE.
The reason is not what one might think.
To cut through a lot of driver design tedium, lets just say that the degree of difficulty or the degree you have conflicting requirements is more or less the bandwidth times acoustic power.
On one hand, the things that govern the hf and lf corners have not changed and are contradicting.
On the other hand, we can make somewhat stronger magnetic gaps and much smaller magnet assemblies; we have the wonders of material science that has given us somewhat better things to make the driver from and in some cases a better understanding leading to better designs.
In the days of the WE driver, power compression was non-existent because the wire would fall off the coil former well before the temp has even doubled the Rdc (about 230C). Old style epoxy dies around 140-160C, modern aramid adhesives can tolerate 350C.
That’s how you can have a 700 Watt 4 inch voice coil, it’s really hot.
That has lead to the high power drivers we have today and as making a high temperature VC is easy while a wall of contradicting requirements impenetrable, the direction has been narrower bandwidth, higher power drivers.
I certainly appreciate strongly the desire to have one acoustic source and for those really keen to do that with one driver I would suggest tinkering with something you can get parts for.
In that light, I would suggest those interested to model a compression driver on a large horn either in AKABAK or McBeans and fiddle with the parts that effect the low corner.
When the horn is large enough to be loading properly, the limit would be suspension compliance. It’s is easy to make a more compliant suspension in the model and in the real world, you cut slits in it or punch little holes in it near the outer perimeter.
This will lower the fundamental resonance of the driver and by making it a wider bandwidth device, you make it easier to bang the dome and lower the maximum broad band acoustic power it can produce.
Since you can start with a much more powerful driver than you need, you can trade some of that off for an extended lf response.
A powerful trick can be used to extend the lf corner further IF the system and horn parameters permit. In an exponential horn, one has a horn with a fixed “high pass” frequency for the impedance transformation part.
For a 30Hz horn, it can expand no faster than doubling it’s area every 24 inches while a 300hz horn is every 2.4 inches etc.
AS one approaches the low corner, the end of the horn loading, the horn becomes increasingly reactive, exhibiting a negative capacitance.
This looks like a mass that increases as the frequency falls and by choosing the right total compliance (suspension and rear volume) the happy condition can be met where the two reactance’s conjugate and that reveals the acoustic load where it would have been obscured by the suspension compliance.
AS what makes flat spl in a large enough horn is radiator velocity a simple circuit illustrates what is happening. The driver is an LC parallel tank circuit driven through a resistor (Rdc) and a tiny inductor (Le).
The acoustic load is an R which is placed across the tank circuit but the horns job is to take that radiation resistance which is a high value and converting it to a low value in line with the Rdc (if Rdc and the radiation resistance are equal then the efficiency is about 50%).
The system bandwidth (assuming the horn is the large) is limited by the parallel L which is the total compliance and the parallel capacitance which is the mass AND the electrical low pass filter formed by the Rdc and Le (winding inductance).
As one can imagine, if one placed a “variable C” in parallel such that it was resonant or conjugating over a bandwidth instead of a single frequency as we are used to , this can extend the region where he radiator has mobility or freedom to move against the load and it’s velocity not restrained by it’s moving mass or restoring force..
In the Tapped horns at work, the efficient designs use the same effect but without a back volume, the total compliance is only the suspension so instead of a big bruiser driver with a small rear volume, the most efficient tend to have a Fs well above the low corner to take advantage of the reactance annulling.
Anyway, by modeling the horn expansion as a hyperbolic expansion, one has some ability to adjust the magnitude of the negative capacitance which when combined with the “adjusted” driver compliance should allow for something pretty different than an un-tinkered system.
The B&C DCM50 is one you might look at, not cheap but already a wide band driver. Fwiw, the bms coax compression drivers can sum into a single source using dsp or into a conventional 2way all pass source passively.
Best,
Tom
I wouldn’t think you would be able to find a modern driver that equals the WE.
The reason is not what one might think.
To cut through a lot of driver design tedium, lets just say that the degree of difficulty or the degree you have conflicting requirements is more or less the bandwidth times acoustic power.
On one hand, the things that govern the hf and lf corners have not changed and are contradicting.
On the other hand, we can make somewhat stronger magnetic gaps and much smaller magnet assemblies; we have the wonders of material science that has given us somewhat better things to make the driver from and in some cases a better understanding leading to better designs.
In the days of the WE driver, power compression was non-existent because the wire would fall off the coil former well before the temp has even doubled the Rdc (about 230C). Old style epoxy dies around 140-160C, modern aramid adhesives can tolerate 350C.
That’s how you can have a 700 Watt 4 inch voice coil, it’s really hot.
That has lead to the high power drivers we have today and as making a high temperature VC is easy while a wall of contradicting requirements impenetrable, the direction has been narrower bandwidth, higher power drivers.
I certainly appreciate strongly the desire to have one acoustic source and for those really keen to do that with one driver I would suggest tinkering with something you can get parts for.
In that light, I would suggest those interested to model a compression driver on a large horn either in AKABAK or McBeans and fiddle with the parts that effect the low corner.
When the horn is large enough to be loading properly, the limit would be suspension compliance. It’s is easy to make a more compliant suspension in the model and in the real world, you cut slits in it or punch little holes in it near the outer perimeter.
This will lower the fundamental resonance of the driver and by making it a wider bandwidth device, you make it easier to bang the dome and lower the maximum broad band acoustic power it can produce.
Since you can start with a much more powerful driver than you need, you can trade some of that off for an extended lf response.
A powerful trick can be used to extend the lf corner further IF the system and horn parameters permit. In an exponential horn, one has a horn with a fixed “high pass” frequency for the impedance transformation part.
For a 30Hz horn, it can expand no faster than doubling it’s area every 24 inches while a 300hz horn is every 2.4 inches etc.
AS one approaches the low corner, the end of the horn loading, the horn becomes increasingly reactive, exhibiting a negative capacitance.
This looks like a mass that increases as the frequency falls and by choosing the right total compliance (suspension and rear volume) the happy condition can be met where the two reactance’s conjugate and that reveals the acoustic load where it would have been obscured by the suspension compliance.
AS what makes flat spl in a large enough horn is radiator velocity a simple circuit illustrates what is happening. The driver is an LC parallel tank circuit driven through a resistor (Rdc) and a tiny inductor (Le).
The acoustic load is an R which is placed across the tank circuit but the horns job is to take that radiation resistance which is a high value and converting it to a low value in line with the Rdc (if Rdc and the radiation resistance are equal then the efficiency is about 50%).
The system bandwidth (assuming the horn is the large) is limited by the parallel L which is the total compliance and the parallel capacitance which is the mass AND the electrical low pass filter formed by the Rdc and Le (winding inductance).
As one can imagine, if one placed a “variable C” in parallel such that it was resonant or conjugating over a bandwidth instead of a single frequency as we are used to , this can extend the region where he radiator has mobility or freedom to move against the load and it’s velocity not restrained by it’s moving mass or restoring force..
In the Tapped horns at work, the efficient designs use the same effect but without a back volume, the total compliance is only the suspension so instead of a big bruiser driver with a small rear volume, the most efficient tend to have a Fs well above the low corner to take advantage of the reactance annulling.
Anyway, by modeling the horn expansion as a hyperbolic expansion, one has some ability to adjust the magnitude of the negative capacitance which when combined with the “adjusted” driver compliance should allow for something pretty different than an un-tinkered system.
The B&C DCM50 is one you might look at, not cheap but already a wide band driver. Fwiw, the bms coax compression drivers can sum into a single source using dsp or into a conventional 2way all pass source passively.
Best,
Tom
As the discussion has progressed, regressed and progressed yet again...I decided to revisit the original youtube video that started this whole discussion.
Originally I just drooled over the technology and then it hit me that the audio in this clip is either terrible or the WE stuff actually sounds as tinny, thin and grating as the audio clip suggests.
Surely Silbatone could record this set-up more professionally and offer some kind of sample that would give a better example of what these speakers are capable of and post them somewhere on the interweb. Others have and are doing it.
During the course of this thread...being the old vintage nut that I am and forced to rely primarily on my subjective senses...I felt some allegiance to the camp that described these speakers as something quite extraordinary from both an engineering and sonic standpoint. After all, those who seemed to express remarkable respect for their performance have had the benefit of actually hearing them and have heard a breathtaking range of other loudspeaker designs (vintage and modern).
So when the objectivist crowd started clamouring for measurement data I understood that this would only provide one part of the formula IMO of what could potentially unravel their magic...perhaps.
But as this thread grew and grew I decided to revisit the Silbatone group and based on their website and Munich catalogue seem long on descriptive hyperbole and short on any hard performance data..old or new...so I'm beginning to understand why some of the more skeptical here are clamouring for some actual measurement data.
Heck, even AR, Quad and other old-school loudspeaker designs managed to publish some hard measurement data as primitive and limited as it was by today's standards...so it seems that Slibatone could do it with all the measurement techniques and equipment now at their dispose.
How exactly, for example, are they producing the new GIP stuff if they're not using measurements of the old gear to work from? Is it simply by breaking down each vintage component bit by bit and then re-tooling the contemporary stuff? I'd rather doubt it.
However, I do think that when Silbatone uses descriptions such as "...the very highest grade of sound reproduction ever achieved..." that measurement data would be one part of whatever yardstick they are utilizing to support that contention.
it's mind boggling to me that someone would evaluate speakers via a recording, and one probably made with some piece of crap iphone. i mean, even if they were recorded with a nagra, it's ridiculous.
they produced hyperbole ad nauseam too; and their graphs often contained a bunch of mullarkey, or at least wishful thinking.
Sorry Dave. Totally forgot to mention all the PSB's you designed mid-2000...
I never said "good for their time" or "superior to their peers". I'm way beyond that, bro. The phrase I'm using is "paradigm shift." I think the WE stuff and the '57 warrant that, but would have a hard time applying it to the AR3 or the 801. I understand why they have their fans, but I'd rather replace all my Coltrane vinyl with MP3's of Diana Krall and Eva Cassidy than have to ever listen to a pair of 801's again (though in fairness to Diana Krall fans I think Jeff Hamilton has exquisite touch and phrasing). No matter where you sit or how much amplification you throw at them they always sound like three drivers in three separate boxes.
Be super keen to know from you what the "proper perspective" is. Wouldn't it be that the WE stuff and the ESL 57 are still sought-after because of their inherent musical qualities which still manage to surpass many contemporary systems, rather than their technical innovations or idiosyncratic flaws?
*Sigh* Back to the AES white papers again...
Okay Dave, I'll call your bluff:
With the decades of accumulated research and linear progressive development since 1928 (WE) and 1957 (ESL 57's), and the infallible double-blind tested speakers made by the mighty Harman International under the guidance of Dr. Toole and Dr. Olive, please put forward your picks for speakers from JBL, Infinity, Harmon/Kardon or Revel that have created a paradigm shift in respect of engineering breakthroughs and cultural impact.
Really, really looking forward to your suggestions.
Last edited: