Haven't gone there yet, I have the massive DVD collection from Herzog's web site with everything (almost).
Funny, after you had read "Thinking, Fast and Slow" I remember you telling me I should be embarrassed by doing the same thing. In fact you said:
Have you changed your mind?
The effect of reading that book wears off if there is no ongoing reinforcement. Humans naturally revert to thinking like humans do. There is no lifetime inoculation from reading a book once, in other words.
On the other hand, research has shown there is something called expert intuition, which can be valuable in fields where "learning" is possible. No so much in economics, say, but more so in engineering.
On the other hand, research has shown there is something called expert intuition, which can be valuable in fields where "learning" is possible. No so much in economics, say, but more so in engineering.
Last edited:
Sounds like a good hint, thank you. One never knows what these "audio" discussions may bring 😉
Sean E. OliveJanuary 16, 2009 at 8:43 PM
Thanks Eric for your compliment and questions. I can only speculate whether like-minded companies might band together and agree on some meaningful loudspeaker specifications. Today, even the companies that subscribe to comprehensive anechoic measurements with high frequency and spatial resolution don't generally publish these curves. The exception to this rule is JBL Professional (see the LRS getfile).
interesting comment from Sean Olive in his blog.
seems like Mr Olive does care about all aspects of loudspeaker performance.
and this
The relevant specifications would include measures of a) frequency response performance that characterize quality of direct,early, late reflected sounds b) a perceptual meaningful measure of nonlinear distortion and maximum SPL c) perhaps spatial quality (related to directivity). I think a) is already possible, c) needs some work, and b) needs even more work.
the fact that he lists " a perceptual meaningful measure of nonlinear distortion and maximum SPL" as the second most important speaker spec to me says it's not a non-issue as the good Dr G suggests...
this does clearly show that while there is agreement as to "no currently accepted "metric"" it's among the top three in terms of performance specs.
Thanks Eric for your compliment and questions. I can only speculate whether like-minded companies might band together and agree on some meaningful loudspeaker specifications. Today, even the companies that subscribe to comprehensive anechoic measurements with high frequency and spatial resolution don't generally publish these curves. The exception to this rule is JBL Professional (see the LRS getfile).
interesting comment from Sean Olive in his blog.
seems like Mr Olive does care about all aspects of loudspeaker performance.
and this
The relevant specifications would include measures of a) frequency response performance that characterize quality of direct,early, late reflected sounds b) a perceptual meaningful measure of nonlinear distortion and maximum SPL c) perhaps spatial quality (related to directivity). I think a) is already possible, c) needs some work, and b) needs even more work.
the fact that he lists " a perceptual meaningful measure of nonlinear distortion and maximum SPL" as the second most important speaker spec to me says it's not a non-issue as the good Dr G suggests...
this does clearly show that while there is agreement as to "no currently accepted "metric"" it's among the top three in terms of performance specs.
I am planing to prepare some intermodulation tests, where one of the tones would be a bass tone. This might be interesting as it will excite large cone excursions and probably some AM modulation.
As a preliminary test, here is a 30+300Hz IMD, with distortion added mathematically, and a pure 30+300Hz twin tone. The distortion profile is the same as in the flute test.
http://pmacura.cz/30+300.zip
This should be audible with good headphones, I am getting a positive ABX result repeatedly.
As a preliminary test, here is a 30+300Hz IMD, with distortion added mathematically, and a pure 30+300Hz twin tone. The distortion profile is the same as in the flute test.
http://pmacura.cz/30+300.zip
This should be audible with good headphones, I am getting a positive ABX result repeatedly.
To me, this clearly makes my point: that we have no meaningful metric of nonlinear distortion. Until we do all this discussion is pretty meaningless. All loudspeakers will "distort" if played loud enough and a metric for when this happens (MaxSPL based on perception, not on damage) would be a great boon to the industry. But talk of the audibility of nonlinear distortion at normal listening levels is just not meaningful if we have no scale on which judgments can be made, no basis for any comparisons. Clearly Sean, and I, would love to have such a metric, but the fact is that we don't and yet we continue to get by nicely without one, so its importance cannot be that great.
Hi there. If you don't have any distortion measurements then what do you measure in production for quality control? Frequency/impulse response and impedance?
I measure very accurate polar response, sometime impulse response, and seldom impedance as it never changes and is mostly just a curiosity. But the polar response was measured on each and every speaker that I made.
In addition there is no way to produce a completely artifact free reference for comparison.
Scott, I don't think that is the issue, I could make a signal that was sufficiently artifact free, but what would I do with it? One could look at a pure tones harmonics and make some valid judgments, but to get to a simple number or curve that was reliable across many different devices would be a massive undertaking and no one is willing to do it. Another indication that the industry does not see it as important enough.
With amplifiers everyone gets excited about THD measurements. But not so with speakers it seems.
Is this because the measures are too "noisy" to be useful?
Is this because the measures are too "noisy" to be useful?
it's the "puppy under the blanket" again.
and yes there is no current "metric" for evaluation but this does nothing to resolve the issue.
i would have thought that you would be eager to resolve the problem but alas it appears the "mantra" is more important then advancing the knowledge base on the subject.
i may be obsessing about a performance parameter that most do not perceive which leaves me wondering why it's clearly audible to me and not to others?
for the purpose of illustration i'll liken a guitar to a loudspeaker in that they can play a range of notes and both can play multiple notes at the same time as in a chord, both will play the notes in tune and at equal level. but in both speakers and guitars , two guitars can play the same chord or note but we are still able to distinguish between the two guitars because the intonation is not the same. loudspeaker "intonation" is the lumped non linear characteristic of the device no?
i still maintain the position that all forms of distortion are to be minimized rather then discounted.
and yes there is no current "metric" for evaluation but this does nothing to resolve the issue.
i would have thought that you would be eager to resolve the problem but alas it appears the "mantra" is more important then advancing the knowledge base on the subject.
i may be obsessing about a performance parameter that most do not perceive which leaves me wondering why it's clearly audible to me and not to others?
for the purpose of illustration i'll liken a guitar to a loudspeaker in that they can play a range of notes and both can play multiple notes at the same time as in a chord, both will play the notes in tune and at equal level. but in both speakers and guitars , two guitars can play the same chord or note but we are still able to distinguish between the two guitars because the intonation is not the same. loudspeaker "intonation" is the lumped non linear characteristic of the device no?
i still maintain the position that all forms of distortion are to be minimized rather then discounted.
I suppose it might be that the THD figures are so poor that no manufacturer wants to bring attention to them. 😀
I'll try to find some online. If so, Pavel may need to up his distortion levels.
Of course, as with amplifiers, the type of distortion is important not just the totality.
From: Celestion 3000 loudspeaker Measurements | Stereophile.com
I'll try to find some online. If so, Pavel may need to up his distortion levels.
Of course, as with amplifiers, the type of distortion is important not just the totality.
From: Celestion 3000 loudspeaker Measurements | Stereophile.com
Attachments
Last edited:
No, that is not correct. It is more correct to say "the lumped linear characteristic of the device"
I still maintain the position that all forms of significant audible distortion are to be minimized. If it is not audible then what is the point of minimizing it?
We should clear up the idea that there are "types" of nonlinear distortion, that THD is a different thing from IMD. They are not, both are symptoms of the same underlying system nonlinearity using special signals. They both derive from the exact same disease.
What is important is to understand how the various orders of the underlying nonlinearity are a factor, because that's where the truth lies. The higher the nonlinear order the greater its audibility will be. In terms of HD that means that the higher the harmonic the more audible the underlying system nonlinearity will be.
i'd back up a step and clarify the definitions of both terms "linear distortion" and "non linear distortion" so that there is less confusion and waffling.
intonation on a guitar is how a luthier minimizes the differences in pitch across the non linear scale of the guitar's neck (i hazard you've never seen a compensated scale neck!(there a bitch to play))
so sorry by my reckoning intonation attempts to the minimize the deviations as in reduce the "non linearity" that would alter the pitch.
aye there's the rub... and i'll use an analogy if i may... a bad wheel bearing!
at highway speeds road noise may mask the sound of a bad bearing until the point of failure! not a good scenario!
so for me the stuff that by current reckoning is not audible and is dismissed does indeed affect the audible that's why is fail to understand the lack of concern.
intonation on a guitar is how a luthier minimizes the differences in pitch across the non linear scale of the guitar's neck (i hazard you've never seen a compensated scale neck!(there a bitch to play))
so sorry by my reckoning intonation attempts to the minimize the deviations as in reduce the "non linearity" that would alter the pitch.
aye there's the rub... and i'll use an analogy if i may... a bad wheel bearing!
at highway speeds road noise may mask the sound of a bad bearing until the point of failure! not a good scenario!
so for me the stuff that by current reckoning is not audible and is dismissed does indeed affect the audible that's why is fail to understand the lack of concern.
- Status
- Not open for further replies.