gedlee said:
Make is loudspeaker amplifier combination that makes distortion at high level related to your measurement on levels that it can't be heard.
And show your theorie in practice with a set that has a normal low distorsion level.
But distorts on noise level as your claiming that's bad.
To show the market a high level distortion isn't always bad. And that a high level distortion system sound natural en better then the one with the nice figures.
To prove the GL factor is a better figure to measure quality than THD and IMD.
Regards, Helmuth
Helmuth
I'm sorry, but I don't follow your comments.
As to proof, just read the reviews of my system on my web site, or any of the reviews by my customers. Not one of them will disagree with my claims. The proof is in the success of the application of the concepts. Beyond that I have no other proof, nor will I ever attempt to provide any. If people want to continue to believe in the old out-dated ideals that's fine, but they will forever go around in a circle pursuing truth when there is none to be had in those obsolete concepts of sound quality.
I'm sorry, but I don't follow your comments.
As to proof, just read the reviews of my system on my web site, or any of the reviews by my customers. Not one of them will disagree with my claims. The proof is in the success of the application of the concepts. Beyond that I have no other proof, nor will I ever attempt to provide any. If people want to continue to believe in the old out-dated ideals that's fine, but they will forever go around in a circle pursuing truth when there is none to be had in those obsolete concepts of sound quality.
I think the problem with acceptance of your new metric is that the industry *doesnt* understand the psychology of hearing. The world is so complex these days that no one can understand it all; it takes a career to understand just one small piece of it all.
Couple that with the gross betrayal of our political and corporate leaders and you have our current environment lacking in knowledge and trust. All we can do, as you are doing, Earl, is try and make our voices heard and stand up for what we believe in.
Anyway, back to the plot.
In your paper, you talk about static nonlinearity implying that there is a dynamic non-linearity. Is this what you mean by memoryless? Douglas Self has done a lot of work on the thermal variations of amplifiers and it strikes me that a quiet passage immediately after a loud passage could suffer bias distortion due to the thermal drift caused by the loud passage. Would you call this a non-linearity with memory?
Couple that with the gross betrayal of our political and corporate leaders and you have our current environment lacking in knowledge and trust. All we can do, as you are doing, Earl, is try and make our voices heard and stand up for what we believe in.
Anyway, back to the plot.
In your paper, you talk about static nonlinearity implying that there is a dynamic non-linearity. Is this what you mean by memoryless? Douglas Self has done a lot of work on the thermal variations of amplifiers and it strikes me that a quiet passage immediately after a loud passage could suffer bias distortion due to the thermal drift caused by the loud passage. Would you call this a non-linearity with memory?
Iain McNeill said:
Not quite right. Nonlinearity CAN depend on frequency - it does in a loudspeaker. When this happens its called "dynamic nonlinearity". For some strange reason when it is frequency independent, its called "memory-less". I didn't make up the terms.
Thermal issues are generally seperate from nonlinearity issues, mostly because the time constants tend to be long enough that they are below the bandwidth of the signal. However, there is some evidence that the thermal constants of small voice coils can be short enough that a true dynamic nonlinearity occurs where the thermal response modulates the signal in a nonlinear manner. But in general the thermal issues can be seperated from the nonlinear ones.
But make no mistake about it, thermal problems are serious. I, like almost everybody else, believed that nonlinear distortion was what we heard as poor sound quality. I personally don't believe that anymore. I believe that in loudspeakers, it is diffraction and thermal effects that dominate what we perceive. Nonlinear distortion is easily designed out and need not be a factor, but diffraction and thermal are NOT so easy to deal with as they are fundametal physics issues that don't simply go away.
In electronics its virtually all about low signal behavior - first watt as Nelson Pass calls it. I completely agree with this. But I don't think that you need mega-buck amps to get this. In fact, due to the better component matching available on a chip amp, they "tend" to be better at low level performance and hence can be great solutions. My system uses chip amps. I've tested them and they are superb in low level performance. Much better than most discrete amps. Of coarse class-A is a no brainer for low level performance, but at a high cost in everything else.
As to the world adopting my metric - I could really care less. As long as my competitors are chasing the wrong things while I do the "right thing", I will always stay ahead of them. SO if they want to design for lower THD when it is totally irrelavent let them. I know better.
gedlee said:
My comment is based on reading your paper about distortion perception.
When I read it and read your comments to tell people about this research. And your own conclusion no one seems to care.
This all does me conclude you want to believe people your right, or to have comment to prove your research was wrong.
When I read our research I can not check your right only by repeating some of your tests. And there is no tool to measure the GM-factor.
Your Summa speaker I know from this forum and your site.
As for the HF driver that is a very low distortion driver at 90dB it will out perform many high end drivers on THD.
I do not know the distortion figures of the (bass)mid driver I gues they are quite good to. Your choises made it a monitor with high dynamic range. It is in fact a qaulity PA monitor.
So you made a low THD distortion speaker compared to hifi speaker both THD measured on 90db 1W/mtr. I assume.
So what is the relation of your low THD speaker with the research that Low THD isn't that important?
(The idea a wave guide is enough and the flare doesnt matter is new to me. The idea to damp the horn with low density foam is very smart.
The idea to place three subs a symmetric I can visualize the standing wave of one sub or two subs can be damped out by placing three subs asymmetrical.)
Regards, Helmuth
On can talk about distortion theoretically all you want.
I was surprised one day when i hooked up a tweeter to a sine wave generator and noticed that isn't didn't sound all that much like a pure single tone. That seemed a bit odd.
I grabbed a Dynaudio D-28 and connected it to the same signal generator and the tone was very pure sounding. Not surprisingly the Dynaudio D-28 was a much better sounding tweeter with music as well.
One of the listening tests that I now do is to listen to the raw drivers connected to a signal generator. It's a good way to weed out the lesser quality drivers.
For the record the D-28 has a harmonic distortion rating of 0.2% and an intermodulation distortion rating of 0.4%
But it is quite true that many of the specifications that audio products are designed for are meaningless.
One only needs to listen and compare an amplifier with a super low harmonic distortion rating with high amounts of loop feedback, to an amplifier with a higher harmonic distortion rating and no loop feedback rating, to realize that steady state harmonic distortion measurements are not applicable to amplifiers with loop feedback.
I can look at the specifications of many raw electronic components and predict which ones will sound better to a degree. But often times important specifications are never published, and other parameters that are important to audio are never measured. I always do listening comparisons.
Scott Novak
I was surprised one day when i hooked up a tweeter to a sine wave generator and noticed that isn't didn't sound all that much like a pure single tone. That seemed a bit odd.
I grabbed a Dynaudio D-28 and connected it to the same signal generator and the tone was very pure sounding. Not surprisingly the Dynaudio D-28 was a much better sounding tweeter with music as well.
One of the listening tests that I now do is to listen to the raw drivers connected to a signal generator. It's a good way to weed out the lesser quality drivers.
For the record the D-28 has a harmonic distortion rating of 0.2% and an intermodulation distortion rating of 0.4%
But it is quite true that many of the specifications that audio products are designed for are meaningless.
One only needs to listen and compare an amplifier with a super low harmonic distortion rating with high amounts of loop feedback, to an amplifier with a higher harmonic distortion rating and no loop feedback rating, to realize that steady state harmonic distortion measurements are not applicable to amplifiers with loop feedback.
I can look at the specifications of many raw electronic components and predict which ones will sound better to a degree. But often times important specifications are never published, and other parameters that are important to audio are never measured. I always do listening comparisons.
Scott Novak
Helmuth said:
THD on a compression driver was measured at 25% and it was not audible when compared to the direct fed signal. Compression drivers ARE NOT low THD devices.
The Summa speakers are not PA monitors, they have almost never been used for that purpose (although they do work quite well). They are almost exclusivley used for Stereo and HT critical listening. They were designed to be audiophile loudspeakers not Pro loudspeakers. I use Pro drivers because they are better drivers.
My point is that I don't care about nonlinear distortion anymore. I don't study it, I don't measure it and I don't look at THD levels in the drivers. Low THD, high THD, it doesn't matter. What matters is polar response, low diffraction and thermal capability. Smooth response is desirable, but ruler flat on any one axis is also irrelavent. THD and IMD are meaningless.
So did I or will I do the tests that you suggest - no. If you want to "prove the validity of THD" then you do it. I'm simply not interested. I did enough tests to make up my mind and if I have not convinced others then so be it. Let them think what they will. I made my case and I'll rest on the evidence presented.
And I am not the only one of this opinion - there is Floyd Toole, Laurie Fincham, Alex Voishvillo and many many other people who spent their lives looking at sound quality.
cbdb said:
The consumer chases what marketing tells them is important. The last thing that marketing wants is truth. They can't market the truth. And what is really a shame is that consumers pay dearly for the manufacturers marketing budget!
gedlee said:
This is not right Earl. I really have to doubt your opinion after this statement.
The B&C DE25 has a THD of 0,1% at 85dB 1W/mtr Measured with horn ME45
and <1,3% at 100dB 1W/mtr! Measured with horn ME45
See " Klang und Ton 03-2004", test of 8 horn tweeters.
Again you are using a extreme low THD HF driver. Also when you compare it to High-end drivers
I agree here and it seems a contradiction to your previous claim (THD level).
I agree this is very important.
Your summa PA or High-end I do not care is a low distortion speaker compared to any speaker measured on the same spl.
So this is in contradiction with your research.
You may not look at the THD but when you look at it at 90dB 1W/mtr I predict a average THD level of 0,5% witch I would rate high ended.
Also if you do not care, the Summa is a low distortion speaker. And so you have done what most builders do, use low THD drivers.
And sure they sound good it doesn't matter what name professional driver , Hifi, High end.
Regards, Helmuth
Helmuth said:
When you phase it that way it completly misses the point. It doesn't matter what the THD is - high or low, it doesn't matter. You keep implying that it does because my speakers sound good and have low THD - that does not prove anything. Michigan has lots of water and no elephants, so they must not like water?
Even if the DE250 is low at 86 dB we tested at something like 120 dB in a plane wave tube - it WAS 25% THD and IT WAS NOT detectable.
You can believe me or not, thats your right, but I have done more work in this area than anyone else and that is my considered opinion.
gedlee said:
The point is, you claim low or high THD isn't important. And after that you are pointing at your speaker programme (the Summa) as an example.
And I say the Summa is a very low distortion speaker that is for sure to say for the HF driver. Compared to any other high quality speaker measured at the same SPL level.
So this is a big contradiction in your story, and my point. We can't hear high THD levels but you build a low THD Summa. Nothing else 😀
By the way. Witch driver does not distort 25% at 120dB? Did the listener also have to endure 120dB at this test?
I did no testing at your level at all.
And appreciate your drive and effort to think different and prove it by testing your theory. I read it very interested.
And will ask critical questions about it if I can.
Helmuth
Helmuth said:
a) you should read the actual journal article, not interpret the powerpoint slides, which are abbreviated
b) Geddes' conclusion was that low or high THD does not correlate with what is actually audible.
The conclusion of the paper is that high THD levels are NOT NECESSARILY audible, because there are other variables to the distortion that affected audibility besides the sum total quantity.
SG
Helmuth,
I think you're missing the point. Dr. Geddes work has shown him that he needs to concentrate on speakers with good polar response rather than low distortion. It just so happens that drivers with good polar response also have low distortion due to using quality drivers but that wasn't his driving motivation. We are fortunate that he is sharing this knowledge with us.
That's not to say that distortion isn't important. For amplifiers it is of the utmost importance because they can generate large amounts of high order harmonic distortion at low listening levels. This is offensive and unacceptable.
If you've ever had the misfortune of working with an echo canceler you'll know that speaker distortion is also very important here as the adaptive filter assumes a linear transfer function.
My interest is in simulating distortion in record/playback/transmission & processing systems for comparative listening tests without having to switch banks of hardware with relays and other suspicious gear.
We have the interest in distortion and Dr Geddes has given us a key piece of the audibility puzzle. It's up to us to do something with it.
I think you're missing the point. Dr. Geddes work has shown him that he needs to concentrate on speakers with good polar response rather than low distortion. It just so happens that drivers with good polar response also have low distortion due to using quality drivers but that wasn't his driving motivation. We are fortunate that he is sharing this knowledge with us.
That's not to say that distortion isn't important. For amplifiers it is of the utmost importance because they can generate large amounts of high order harmonic distortion at low listening levels. This is offensive and unacceptable.
If you've ever had the misfortune of working with an echo canceler you'll know that speaker distortion is also very important here as the adaptive filter assumes a linear transfer function.
My interest is in simulating distortion in record/playback/transmission & processing systems for comparative listening tests without having to switch banks of hardware with relays and other suspicious gear.
We have the interest in distortion and Dr Geddes has given us a key piece of the audibility puzzle. It's up to us to do something with it.
Helmuth said:
And there are no elephants in Michigan - proof positive that elephants don't like water.
There IS NO contradiction. Only coincidence.
Intrigued by the I/O curve idea and seeing as I had a bad EMU 1616 which, for some undiagnosed reason, was putting out 1%THD, I thought I take a look at what could be done.
I started with a 1KHz sine generated in Adobe Audition, plotting output vs input I got the following plot:
I started with a 1KHz sine generated in Adobe Audition, plotting output vs input I got the following plot:
Attachments
I was really encouraged by this strange picture because it showed the latency in the system as a phase shift and the start-up artifacts showed up as wild deviations from an ideal lissajous figure.
Scott Novak said:
Yes it's as simple as that. My experience is exactly the same both with tweeters and with mids/basses. I almost never get close the significant thermal compression but still hear this and with similar baffel layouts and dispersion. Main changing parameter has been nonlinear distortion and I always prefer the lower.
/Peter
Iain McNeill said:
Sort of, but I said "rough idea". Masking is poor for IMD and high levels of HD means high levels of IMD
Yes it has. If a driver sounds like you're feeding it a triangular or square wave mixed with a sine when you're actually are feeding it a sine you can be pretty sure that for music it will distort audibly.
/Peter
- Status
- Not open for further replies.