Except not vented, so smaller cabinet dimensions for the paired 15" woofers. VLF taken care of by self-powered subwoofers at different locations.
Regarding Dr. Geddes' thoughtful comments about the order of the distortion, that echoes the suggestion of the D.E.L. Shorter and the BBC Research Labs in the Fifties to weight harmonics by the square or even the cube of the order to better reflect audibility. It was not computationally realistic back then, but not a problem today.
In the vacuum-tube world, some collectors obsess about the measured THD of this or that tube, but it hardly matters. Triode distortion is almost entirely 2nd-harmonic, and very difficult to hear. However ... pentodes have generous amounts of higher-order distortion, particularly when operated in Class AB, so the 20 dB of feedback in a typical mid-Fifties Golden Age amplifier has only a modest effect in reducing perceived distortion. Thus, the mushy/grainy/jukebox sound of these old amplifiers.
Regarding Dr. Geddes' thoughtful comments about the order of the distortion, that echoes the suggestion of the D.E.L. Shorter and the BBC Research Labs in the Fifties to weight harmonics by the square or even the cube of the order to better reflect audibility. It was not computationally realistic back then, but not a problem today.
In the vacuum-tube world, some collectors obsess about the measured THD of this or that tube, but it hardly matters. Triode distortion is almost entirely 2nd-harmonic, and very difficult to hear. However ... pentodes have generous amounts of higher-order distortion, particularly when operated in Class AB, so the 20 dB of feedback in a typical mid-Fifties Golden Age amplifier has only a modest effect in reducing perceived distortion. Thus, the mushy/grainy/jukebox sound of these old amplifiers.
Last edited:
You, me and Earl, everybody, all hear some of "these things." I don't think that's disputable. What we mostly argue about is the cause and sometimes the efficacy of solutions. I don't see how these disputes can be resolved without measurements both physical and psychoacoustical.
And, sometimes the problem for a confused person such as myself, does not sit under a normal distribution. For example, I'm very sensitive to IM around 3 to 5.5 Khz - I can even hear it in live performances if I'm close to musicians - it's only my right ear but it's very unpleasant. I used to think the problem was with the gear, and sometimes it is, but fundamentally, the distortion I was hearing, that others were not, was listener generated.
But once I had idea, derived from the live performance experience, the problem might be a combination of my physiology and bummy speakers there was measuring to do, so I could know, not just have an idea.
Not everything - there is science and measurements behind using foam in a waveguide.
Last edited:
Nice posts on a good topic by Earl and Bill, picked up by Lynn.
When an SET amp is properly built (many are not) the distortion will not be audible below clipping. A regular fall off of the harmonics is masked by the ear.
I don't know too much about harmonic audibility in speaker cones, except that it's much easier to hear on pure tones than in complex sounds. I can imagine that certain spectra could compliment each other as Bill says, as they can with pentodes and triodes. Maybe with amps and speakers, too.
Well worth looking at I think.
When an SET amp is properly built (many are not) the distortion will not be audible below clipping. A regular fall off of the harmonics is masked by the ear.
I don't know too much about harmonic audibility in speaker cones, except that it's much easier to hear on pure tones than in complex sounds. I can imagine that certain spectra could compliment each other as Bill says, as they can with pentodes and triodes. Maybe with amps and speakers, too.
Well worth looking at I think.
I've seen some measurements here (I believe in the NCore thread) that indicated lower midrange distortion with high output Z amps compared to very high-performing low output impedance amps. This seemed to be a result of the speaker distortion actually lowering, not the amp. Bass distortion remained better with the NCore compared to what I believe was a pass labs from the f-series.
This is the effect of (oh so overlooked) current driving. A fellow Finn Esa Meriläinen has done a great research: http://www.current-drive.info/
My own measurements can be found here: http://www.diyaudio.com/forums/solid-state/250272-current-drive-loudspeakers-22.html#post3838473
Very easy to test with a series resistor how the odd-order harmonics get lower and lower as you increase the resistor's value.
Is also believe current driving to be the reason why passive xo'ed dynamic speaker sound better than active xo'ed. Passive crossover creates current driving situations, decouples the driver from the driving amplifier. Especially compression drivers can be current driven in crossover since they usually need quite much attenuation.
This is also behind the tube amplifiers preferred sound, as one of the contributors. SET amps without global feedback have output Z of many ohms.
Last edited:
This is interesting - I have never heard that happen even with massive drivers loaded in bass horns. You must be talking about the mid and small format compression drivers?
Correct, just to point out that one should not only mix & match frequency response, efficiency, dispersion but also balance dynamically.
Trying to remember it could also have been the PD121 (it is over 10 years ago and I have heard and used many smaller and medium PA speakers my friend produces), but it is not important exactly which driver of the two it was (I can easily ask my friend when I want to know). The important aspect is that it is hard to predict from a spec sheet for any driver (woofer, bas-midrange, compression driver, ribbon tweeter, dome tweeter etc.)
Current drive does definitley help in reducing driver harmonic distortion.
But I guess this driver THD reduction is only partly the reason for some people prefering single ended class A.
I guess the bigger part is due to the fact that the distortion caused by these amps is not only of low and even order but it is getting lower with lower drive levels.
While the crossover distortion of AB amps for instance is increasing with lowering drive levels - which might not be a good idea in terms of masking fine details.
Regards
Charles
But I guess this driver THD reduction is only partly the reason for some people prefering single ended class A.
I guess the bigger part is due to the fact that the distortion caused by these amps is not only of low and even order but it is getting lower with lower drive levels.
While the crossover distortion of AB amps for instance is increasing with lowering drive levels - which might not be a good idea in terms of masking fine details.
Regards
Charles
Peter. This is spot on. From two dome tweeters that seemed to spec almost the same, sounded very different with excessive sibilance in one. The tweeters were Al versus titanium domed.
Specifying audio performance seems to still be in its infancy, along with dynamic circuit performance because it is not possible to design from specification without considerable practical adjustment and changes.
Last edited:
About harmonic spectra and SET with the speaker, does the output traffo make "better" the output harmonics of the Tube behind ?
If now instead we have SETransistor amp like a Nemis with huge caps bank for current (no active regulator) and an output traffo like the first design Hiraga made, could we have Something near the inputs above ?
(I don't remember the Z output of him design)... but remember Pass said thi design was better with the outputtraffo than without (I assume in their tries to copy the sound signature of tubes with transistors)...
If now instead we have SETransistor amp like a Nemis with huge caps bank for current (no active regulator) and an output traffo like the first design Hiraga made, could we have Something near the inputs above ?
(I don't remember the Z output of him design)... but remember Pass said thi design was better with the outputtraffo than without (I assume in their tries to copy the sound signature of tubes with transistors)...
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.
This is not a debate; we're talking about a mechanism that is either supported by the physics or it's not. I'll give you the benefit of the doubt and try to better understand what you're thinking; are you saying that the air molecules in front of the tweeter dome present less of a resistance to the motion of the dome since they're not bounded inside a compression chamber? Is this why it's supposedly a little easier for the dome to move "very small increments"?
As to the lower distortion from a constant current amp: If one looks at the nonlinear diffEQ for a transducer it is clear that the inductance nonlinearity is nulled out in a constant current device. This is dominantly a mid-range effect. BUT, in a loudspeaker that has a flux modulation ring, this nonlinearity is also nulled out and as such the constant current amp will not have an effect on this nonlinearity in that case.
It would be exceedingly important to understand nonlinearity in loudspeakers in detail if it was an audible effect. This is why I spent nearly 10 years studying it - learning nonlinear dynamics, modeling transducers, etc. It was only when I did the audibility studies that I found that these effects were essentially not audible that I stopped study them (I should note that ALL studies of these effects have arrived at this same conclusion to the point where Toole does not even mention them in his book.) To me, these days, THD in a loudspeaker is simply not worth worrying about. It does not really even enter into the big picture of what makes a great loudspeaker system.
Of course, it is, as I said before, an audible effect in electronics and some people even like it. But since I am not a believer in the audibility of amplifiers - these effects can be made to be insignificant even though that is not always the case - I am not too interested in nonlinearity any more.
It would be exceedingly important to understand nonlinearity in loudspeakers in detail if it was an audible effect. This is why I spent nearly 10 years studying it - learning nonlinear dynamics, modeling transducers, etc. It was only when I did the audibility studies that I found that these effects were essentially not audible that I stopped study them (I should note that ALL studies of these effects have arrived at this same conclusion to the point where Toole does not even mention them in his book.) To me, these days, THD in a loudspeaker is simply not worth worrying about. It does not really even enter into the big picture of what makes a great loudspeaker system.
Of course, it is, as I said before, an audible effect in electronics and some people even like it. But since I am not a believer in the audibility of amplifiers - these effects can be made to be insignificant even though that is not always the case - I am not too interested in nonlinearity any more.
Last edited:
I too found this claim to be beyond my understanding, but I let it pass. I am curious, like you, what was meant here.
Earl,
Trying to follow your comment on the inaudibility of a nonlinear loudspeaker are you saying that if we intentionally made a speaker with an offset voicecoil that it would be a non issue? Let's say we take it to the extreme and placed a long coil in a gap with half the coil outside the gap and the other half in an equal length gap to the 1/2 coil. Are you saying that this asymmetric nonlinearity would be inaudible or am I just taking this statement to far?
Or are you saying that the asymmetric motion around the normal resting point being equal in both the positive and negative motion would be inaudible which I can imagine would be more the case?
I think this follows in my statement that the Doppler shift caused by the motion of a long excursion driver is also not really audible though in theory it can be shown to exist and measured. This is the argument so often used for using large bass drivers as it is argued that the Doppler shift in these short distances can be detected by the human ear. Front to back detection of distance over these very short distances would seem to not have a way for our ears to detect it if they are directly in front of our ears with no movement side to side. Kgrlee has told me that in DBLT that Doppler shift of small drivers was not detectable, that it is another audio myth.
This is not saying that high distortion is not detected but that this has nothing to do with the front to back Doppler shift itself. So a small driver with the same level of distortion would not sound any different than a larger driver of the same distortion level and harmonic structure.
Trying to follow your comment on the inaudibility of a nonlinear loudspeaker are you saying that if we intentionally made a speaker with an offset voicecoil that it would be a non issue? Let's say we take it to the extreme and placed a long coil in a gap with half the coil outside the gap and the other half in an equal length gap to the 1/2 coil. Are you saying that this asymmetric nonlinearity would be inaudible or am I just taking this statement to far?
Or are you saying that the asymmetric motion around the normal resting point being equal in both the positive and negative motion would be inaudible which I can imagine would be more the case?
I think this follows in my statement that the Doppler shift caused by the motion of a long excursion driver is also not really audible though in theory it can be shown to exist and measured. This is the argument so often used for using large bass drivers as it is argued that the Doppler shift in these short distances can be detected by the human ear. Front to back detection of distance over these very short distances would seem to not have a way for our ears to detect it if they are directly in front of our ears with no movement side to side. Kgrlee has told me that in DBLT that Doppler shift of small drivers was not detectable, that it is another audio myth.
This is not saying that high distortion is not detected but that this has nothing to do with the front to back Doppler shift itself. So a small driver with the same level of distortion would not sound any different than a larger driver of the same distortion level and harmonic structure.
It is interesting how we have different ways of looking into the situation...
Often someone reports a phenomenon with a suggested reason for cause. Often I disagree with the reason but my focus has always been on the phenomenon, not the reason. This is way more challenging to the mind/brain.
Here is the situation: John had an experience [may be several times] where using larger bass drivers brought better experience than using smaller bass drivers. Knowledgeable or not, it is a human nature to find an explanation for any phenomenon. And it happens that he picked Doppler effect for the explanation...
Jack, an objectivist, examined the Doppler issue and came up with a conclusion that Doppler effect is a myth. And he might conclude that there is no difference between small driver and big driver, provided that bla bla bla, which can be dangerously misleading.
What is missing here is an understanding of the real situation which created the phenomenon in the first place, such that useful precautions or actions regarding the issue can be achieved.
And quite unfortunately, using random ears to prove the point, in case of a null result, is a very big mistake imho. Because I know that many people have bad ears.
ADD:
I think there was a story from someone on this website, where you can dismiss a crowd by playing a very high frequency at a very high SPL. Can the crowd hear this frequency [imagine this is the blind test]? No. But can the frequency have an effect to the crowd? I think possibly. This is what I mean that most people have bad ears. Because there are different ways we can "hear" or "feel" sound.
Last edited:
Jay,
If you happen to be one who dismisses DBLT as inconclusive or irrelevant to audio testing then there is no real testing in any real sense that can ever apply. There are truly things that we have not seemed to come up with a test to quantify, at the same time there are tests that find differences that are measurable that have been found to be irrelevant to our hearing. I am not saying that you have to only go by the current tests, objective, as the only answer and that there are not subjective differences between speakers that measure nearly identical. I think it is a mix of both, you can't take one to discount the other and vice versa. But there are phenomena that are attributed to some objective science that just don't hold up once you bring in the anatomical thing called ears into the equation. No two people ever seem to describe a sound in the same way using words, does that mean that we can't discuss the subject?
If you happen to be one who dismisses DBLT as inconclusive or irrelevant to audio testing then there is no real testing in any real sense that can ever apply. There are truly things that we have not seemed to come up with a test to quantify, at the same time there are tests that find differences that are measurable that have been found to be irrelevant to our hearing. I am not saying that you have to only go by the current tests, objective, as the only answer and that there are not subjective differences between speakers that measure nearly identical. I think it is a mix of both, you can't take one to discount the other and vice versa. But there are phenomena that are attributed to some objective science that just don't hold up once you bring in the anatomical thing called ears into the equation. No two people ever seem to describe a sound in the same way using words, does that mean that we can't discuss the subject?
Could I design a driver that had very high levels of audible nonlinearity - absolutely. Could I design one in which the nonlinearity was inaudible until pushed to its limits (which I can define,) absolutely. If you are using drivers with audible nonlinearities then use another manufacturer - one that understands the subject. Like B&C - and probably JBL, and perhaps some others. But don't ASSUME that all driver designers do understand this problem - like most Chinese engineers (I should know, the few that do know this stuff learned it from me.)
It is rather clear what you say about the lower resistance, but consider the dome is directly thrusting the air molecules forward. In a CD the diaphragm is squirting the turbulating air, compressing it so it converts frictional energy to heat, most of which is lost and then passes through an annular or radial phase plug, losing more heat in transit, then with a falling and cooling pressure gradient. This will significantly affect small diaphragm displacements, and do it non linearly over the range of the frequencies we want. The highest frequencies at low levels will be attenutated more and more, subject to the greater viscous stiffness losses of the CD tweeter under comparison, for say 2" aluminium
diaphragm. The diaphragm in the CD has to be of a disprortionately
heavier construction to over come the greater pressure and heat while matching the stiffness and fidelity of 1" direct Al dome tweeter. All these losses are energy and information loss, or converted into unwanted distortion.
If you are listening to quieter passages of complex music with fine detail,
and set the drivers to the same low but typical soft passage sound level , they may be close in detail, but the dome direct dome will be the winner delivering more accurate audible microdetail. CD,s also inherently fall away in the treble due to cancellation in the throat. The subtle inner detail is more affected for being generally of higher frequency content.
So there is also quite a package of small losses which adds up against 1" Al dome which have a lower level of these losses.
The debate excludes any horn or waveguide which is another matter.
Last edited:
What this means is that if you do not have three criteria met; 1) a scientific principle to define a phenomena; 2) measurements to confirm its existence; 3) blind, multiple subject listening tests to confirm its audibility; then it probably isn't worth talking about, because 99% of the time the "phenomena" will turn out to be irrelevant at best or completely bogus.