It's important to remember that mediocrity is the norm in most any human endeavour. Academia is no exception, although its denizens may well be affected with the greatest amount of pretentiousness
When the CTO of a certain automotive company learned that said was about to be acquired by Harman, and that Sid Harman had a PhD (in Social Psychology from Union, iirc, something which required life experience, an essay, and a hefty check), he quickly set about getting his own advanced degree. His name was already on a plethora of publications, having managed a certain large laboratory for a number of years, although I believe he was rarely the lead author. The non-technical degree was swiftly obtained, although the dissertation is not available. But at least the two characters could address each other as Dr.
This is why books (in whatever physical form) will never completely go away, and as well why Wikipedia can be so valuable. There will always be a need for concentration and a comprehensive point of view, assembled by people who do know the given field intimately.
Unfortunately, good technical books are of necessity a labor of love and perhaps pursuit of recognition and validation, but their authoring generally not a profitable occupation. If you are very clever you might write a textbook that is widely adopted for introductory courses. The late George O. Abell worked hard at its writing and updating, but The Exploration of the Universe did bring in of order $100k/year --- back when that was a pretty large amount of money.
Needless to say, he already had a secure position. Textbooks won't impress committees that award or deny promotions.
Now that there is a quiet moment, I would like to explain what I am talking about when it comes to circuit design. AND circuit design is not ALL that electronic engineering encompasses. I am a circuit designer, primarily.
I have found that INVENTIVE circuit design does NOT necessarily involve fully understanding transistor action, or the physics behind active devices. It fact, IF you know too much, you just 'might' miss trying something that you are SURE will not work in advance, because of your foreknowledge of a device's limitations.
For example, I designed the Levinson JC-1 pre-preamp when I was just barley able to operate as an independent designer. If you look closely at my patented circuit, you will find that SOME transistors will NOT work with it, yet many other popular parts will, AND in 1973, we were able to introduce a 0.4V/rt-Hz pre-preamp that didn't cost a fortune. If I were to attempt to design something so radical, today, I doubt that I could get away with attempting it.
I have found that INVENTIVE circuit design does NOT necessarily involve fully understanding transistor action, or the physics behind active devices. It fact, IF you know too much, you just 'might' miss trying something that you are SURE will not work in advance, because of your foreknowledge of a device's limitations.
For example, I designed the Levinson JC-1 pre-preamp when I was just barley able to operate as an independent designer. If you look closely at my patented circuit, you will find that SOME transistors will NOT work with it, yet many other popular parts will, AND in 1973, we were able to introduce a 0.4V/rt-Hz pre-preamp that didn't cost a fortune. If I were to attempt to design something so radical, today, I doubt that I could get away with attempting it.
What worked and what didn't work was determined by what was going on deep inside, quasi-sat, etc. The deep understanding would have helped you in this case.
No. That IS my point about thd and fft. Scott got it on the first take... reread his answere. The ear is sensitive to waveform shape, amoungst other things, and knowing the average level of a harmonic does not correlelate to what might have been heard because of this fact. just a point to be made about measuring harmonic levels in design of Audio amps. For non-audio, it is better suited and offers a direct correlation to what the results will be. -RNM
also includes examples that appear to be direct refutations of the "waveform shape is more important" rubric
like the relative inaudibility of ms all pass phase shift networks, appearance of multiple spatially separated dynamic drivers, LR4 phase inverting crossovers in "high end" speakers...
psychoacoustics is complex, neither time nor frequency domain plots of music waveforms read by human eyeballs, interpreted by the small set of simple heuristics you can keep in conscious working memory will let you say what or if every feature you can point to on either time or frequency plots will sound like
Last edited:
No need to... I made the straw man to bring up the point. I use FFt/THd/MLSSA/IM et al. But, I am leaning to finding correlations. I think my input has demonstarated that.
On this forum and this DIY site there are lots of fft graphs of someone's favorite design. So, they go back and forth on listening and fft data or IM or what-ever. Without correlation.
Is this nieve to ask you/others here to try and think about doing tests that do correlate better or even easier? If we dont or cant, who will? Are we doomed to this situation?
Is anyone here capable of designing test equipment that can correlate better? What would we need to do that? Or, tests that do with what we have? Does anyone even care?
Thx, Dick
measures that correlate with listening impressions will require someone to start presenting Psychoacoustic Science informed, controlled, Blinded listening tests with positive results – correlating with circuit or waveform features
so far the very 1st step (precondition really) appears to be a insurmountable hurdle for the "subjectivist" Audiophile community
we do have the GedLee Metric - rarely brought up but actually has some human test data: http://www.gedlee.com/distortion_perception.htm
so far the very 1st step (precondition really) appears to be a insurmountable hurdle for the "subjectivist" Audiophile community
we do have the GedLee Metric - rarely brought up but actually has some human test data: http://www.gedlee.com/distortion_perception.htm
Last edited:
One approach: to predict all possible flaws of all possible topologies.
Another approach: to have exact model of sound perception.
While it is impossible (see Sound Quality and Measurements thread in Lounge) design of audio equipment is an art that uses science and technology.
Listening tests are very boring for the participants, blinded even more so; pyschoacoutics then "says" the results are pretty meaningless because the people being tested give up being interested in what's going on ... the human being has got better things to do then listen to repetitive "noise".
I would like to see ABXs testing methodology compared, where group 1 gets nothing for their efforts, group 2 are promised a nice afternoon tea after the session, and group 3 gets a thousand dollars shared between them if they can distinguish different scenarios at a good enough level. But, some might say, that is irrelevant to scientific understanding ...
Frank
And indeed it is. The last auditory DBT I participated in was one of audibility of midrange phase shift (all pass). No rewards, wasn't easy, but some of us are more motivated than others by sheer curiosity about what we can and can't hear. The unmotivated ones just make excuses, the motivated ones actually learn something, but that's how it always is, isn't it?
My usual tests are more than blind: when people show by their reactions that their imagination had been fooled no need for any other methodology. Either somebody is scared, or say something about sounds from outside, or ask who plays piano inside, and so on... People even do not know that they participate in tests. Even I don't know, I just collect such spontaneous reactions.
It sure wont be easy and could take years. My hats off to F.Toole for taking on what he did. He found sensitivities we thought only Golden ears could hear. So without test equipment nor anyone willing to do the long term, full time research (anyone need a PHd thesis?), I have to use the idea that if a big enough number of people hear something, the chances are its true. If it is completely nuts, then forget it of course.
If one took in such info along with good engineering practices etal, and folded in as much of both into a product.... the price might go up... but some improvement will be made over time.
This is what I see JC as having done. A whole life time -- wow -- on just adding in or folding a little here and a little there as he gets input until the product seems like magic. John says he wont tell all. Guess we'll have to raid his place when he dies and read his notes. JC - take more notes.
But there are some things we will get wrong and fold that in too. But odds are if enough people say something matters, then it does and you wont be adding in unnecessary expense. That does happen a lot in high-end.
I am going to see if others elsewhere have the desire to be challenged to find test measurments that do correlate better than what we have now. Thx for all input... we got there.
-Dick Marsh
BTW GedDLee and Toole are over on Facebook with a group you can join in and get thier candid thoughts and ask questions.
If one took in such info along with good engineering practices etal, and folded in as much of both into a product.... the price might go up... but some improvement will be made over time.
This is what I see JC as having done. A whole life time -- wow -- on just adding in or folding a little here and a little there as he gets input until the product seems like magic. John says he wont tell all. Guess we'll have to raid his place when he dies and read his notes. JC - take more notes.
But there are some things we will get wrong and fold that in too. But odds are if enough people say something matters, then it does and you wont be adding in unnecessary expense. That does happen a lot in high-end.
I am going to see if others elsewhere have the desire to be challenged to find test measurments that do correlate better than what we have now. Thx for all input... we got there.
-Dick Marsh
BTW GedDLee and Toole are over on Facebook with a group you can join in and get thier candid thoughts and ask questions.
Last edited:
The closer you get, the more all the little somethings do start to really figure, it can drive a sane man
... best to just accept that what you have is good enough sometimes ... Frank
- Status
- Not open for further replies.