I don't know what the problem is here, but I also have designed the electronics for condenser microphones as well as the rest of the chain. My first effort was making measurement microphone electronics to match up with a 1" B&K capsule. We made two of these back in 1971. One still is with me today, when I need it. Later, I modified the Nagra based electronics for the B&K 1/2" capsules (about 10) with higher value input resistors, and lowered the noise 10dB for the GD, and added variable sensitivity by controlling the 200V bias supply. Then, I made some custom microphone electronics for some German capsules for a customer, in the early 1980's.
I still have a couple of 1/2" B&K capsules, as well as a couple of Sony 500 capsules, but I have never had the need to build the electronics for them. Most microphone electronics for these capsules, 'suck' compared to what I design for them, except for the B&K electronics, which is usually, just OK. Of course, I am a perfectionist, so my criticisms come easily. Of course I do also make standard microphone electronics for moving magnet and ribbon microphones as well, but that is another set of problems.
I still have a couple of 1/2" B&K capsules, as well as a couple of Sony 500 capsules, but I have never had the need to build the electronics for them. Most microphone electronics for these capsules, 'suck' compared to what I design for them, except for the B&K electronics, which is usually, just OK. Of course, I am a perfectionist, so my criticisms come easily. Of course I do also make standard microphone electronics for moving magnet and ribbon microphones as well, but that is another set of problems.
The biggest issue with original U47's are the VF14 tube being unobtanium.
The one in your link is a new version and uses a different tube which is actually a glass tube inside that metal envelope.
Also the originals were run in starved heater mode, some info:
http://www.wagner-microphones.com/compare inside3.htm
Transformer was dual bobbin, Pi wound and I don't think it had the usual 80/20 core material.
http://tab-funkenwerk.com/id79.html
http://www.tab-funkenwerk.com/id58.html
Local guy (G Wagner), makes one of the best current re-issues of a U47.
http://www.wagner-microphones.com/
T
Sorry, I read cbcb's criticism wrong. I am NOT a recording engineer. I just design electronics for recording engineers. I would not know my way around a modern, or even a 50 year old recording studio. I only visited them, sometimes, to participate in some production.
However, some recording engineers want more than a typical studio provides, and that's when they bring me in.
However, some recording engineers want more than a typical studio provides, and that's when they bring me in.
Last edited:
Ella and Louis is nice too, with Oscar Peterson, Buddy Rich, Herb Ellis, and Ray Brown.
https://www.allmusic.com/album/ella-and-louis-mw0000191819
Last edited:
OK, that explains why new physics and things like directional cables are not discussed openly; they are all trade secrets. How dumb of me not to think of such a simple explanation.john curl said:
Then there are military secrets; some are advanced science, some are a scam, some are merely nonsense.
In olden days it could have been so you didn't spill the beans on all the phone tapping going on, if you came across some of it or overheard colleagues talking about it.scottjoplin said:
Does this not represent a very strong bias by you even to the extent that you will misuse well know & established definitions of "false negative"?
I certainly don't hold the position that all blind tests are flawed & again this is a miss-statement of all that has been discussed so far.
I do believe however that we need some better measurements, as discussed recently here.
It follows from the same, already cited several times in the other threads, comparisons between ABX and other test protocols. When testing the same sensory difference the proportion of correct answers (poc) was found to be significantly lower in ABX than in the other tests.
So, if the statistical analysis is based solely on the number of correct responses (accumulated as test statistic used) then it gets more difficult to obtain a correct positive result.
Which obviously is true, and Leventhal´s analysis was indeed universal in this regard. But, what it makes our discussions in these threads imo quite often so difficult is a premise - usually not outspoken - but assumed to be true, namely that listening under test conditions must be (in a miraculous way) the same as under "normal/casual" conditions, if not that it even must be better (because of the missing "sight confounder").
Leventhal´s examples use the binomial test and the effect size is therefore expressed in proportion of correct (over time), as you have to assume an effect size for any power analysis.
But there is a distinction between the "real" effect size and the effect size under test conditions.
That is one of the reasons, why i consider the "gorilla" example as good illustration of this distinction; the "gorilla" represents imo an effect that is not small, but can be nevertheless be transformed into a "small effect under test conditions.
See above; so it indeed reflects the reality of any empirical experiment, and it might have been used to discredit any "blind test attempt" (not only the ABX variant) but otoh especially Clark and others tried hard to discredit Leventhal´s analyis as unjustified "audiophile nonsense", although as we know Leventhal simply did report something that Clark et al. should have been aware of, but apparently weren´t.
Of course; the first response from Shanefiled and Clark et al. i´ve seen was there Author´s response to Leventhal´s critic (both in the JAES). According to the rules of such journals it was a reasonable one - failing for obvious reason when trying to reject the critic - but already presenting the ad hoc argument that any of these alledegly small differences would be of no practical relevance, the argument brought up by Dan Shanefield.
The response in the letter section of Stereophile to Leventhal´s critic by Clark (and Nousaine it think, Shanefield did not participate) not restraint by the strict Journal rules was far less polite if not overly aggressive and tried to rebutt Leventhal´s analysis with questionable if not absurd (from a statistical view point) arguments.
Reading that i was quite shocked by the missing professional attitude and began to question their motives, as it seemed that they were not so much interested in finding the truth.
The articles i´ve cited were not about synthetical but real differences/positives and showed the reported differences under real experimental conditions.
In association with other analysis attempts there were additional simulations done wrt predictions following from modelling the internal judgement processes. I have to do dig some deeper in my library for that......
Surely, but that brings us back to the (hidden) premise i´ve mentioned above. We have to remember that in most other fields with "controlled blind experiments" there usually is a variable assessable by conventional measuring device, so in these experiments with humans there is objective measurement (still variable between individuals though) of an effect and in addition a response given by subjective "rating" from the humans.
In listening tests this objective part is missing (Oohashi et al. tried to reach something comparable by incorporating PET scans and EEG; recently some researcher tried fMRI to assess listeners response to different reverberation times for concert hall acoustics) and all we have is the subjective response by the participants that might show mainly impact by the test condition than response to the alteration of the independent variable.
Agreed on the sieve property, but in reality that doesn´t help in most cases.
Let me mention again Shanefield´s ad hoc argument of missing practical relevance. It seems plausible, but the scientific approach requires to check if it is true in reality. I have never heard about any attempt by the ABX-audio test crew to do so. But i´ve to mention that Arny Krueger (on his old pcabx page) did very well in mentioning the need to do training for being really prepared and even cited the ITU-Recommendations.
Have you ever seen Martin Kantola's Panphonic Mic project? Years ago he built one with 4 original capsules and NOS VF14's. IIRC even then the VF14's could fetch $1500.
The factors above all discredit any short testing regimen. Forget about variability between different humans, my opinion after testing many is there is very little correlation between most people for anything more than gross differences; but even within a single test subject variability is significant. Our perception varies second by second as muscles tense in our jaw, in our inner ear, as what we are looking at changes, as our minds wander from one thing to another, as we feel hungry, thirsty, have minor sinus and other ailments, add to this expectation bias which is subtle and devilish to train away except by repetition...Longer testing sessions over a few days or longer average out these factors as much as possible.
Except for gross sonic differences, I believe any of these tests really only test the individuals, not the equipment. Of course there is value in that result, but not necessarily as applied to subtle equipment differences.
Cheers,
Howie
Interesting (depressing for me) coincidence that you should mention that, I've recently begun to notice that sometimes my tinnitus has started to vary, usually for the worse, when I chew
Well, imo, if gorillas are not detected it is a small effect (yes, the animal is quite large...) - so small that gorillas does in fact not matter. Then some might think that gorillas are a big deal once they learn that they are there (=sighted
). But there is the whole trick - the hypotesis on trail is "gorillas are not detectable" (or; electrolytic capacitors sound no different from film capacitors) - if they cant, they aren't.
So you can watch basketball with gorillas your whole life and don't suffer at all i.e. having the same experience as the ones watching basketball without gorillas.
Gorilla-less basketball tickets are also much cheaper
Ignorance can be a bliss.
//
Interesting (depressing for me) coincidence that you should mention that, I've recently begun to notice that sometimes my tinnitus has started to vary, usually for the worse, when I chew
Yes, I can tell when I wake up if there was a teeth clenching event (even if a tooth was not broken). My dentist showed me a study where subjects wore strain gauges at night, you honestly would not believe the pressure numbers.
As a scientist you know that plausible hypothesises are still hypothesises and have to be examined in a sufficiently way.
One way to do so is the comparison of poc (proportion of correct trials) with stimuli that are known (either known to be perceptible or known to be identical). Something that was done routinely within the experiments for SDT (Signal Detection Theory) since the ~1940/50s.
If you analyze well documented listening tests you´ll see that for example the poc in case of presenting identical stimuli twice might be as low as 20% (under controlled "blind" listening test conditions) in multidimensional evaluations .
As said before there is a risk of being flawed due to the "knowing about" effect, but it depends on a lot of variables. I think i´ve proposed it before in other threads, it helps a lot to do some experiments yourself (means conducting "blind" tests with other people) to get a better fealing what is really happening under such "unusual" conditions.
Test some people that don´t know neither about "high end" at all nor about specific brands. If you test for amplifier differences you´ll notice that the usual black box design of amplifiers does not attract them and that it often does not make a difference if one black box is named "......" and the other "......" .
That it could be quite different if doing tests with "high end" afficionados should be mentioned.
As usual it depends. If you know about the pitfalls (and know about the countermeasures) of a method, it still can be useful.
If you don´t care (reason might be ignorance or arrogance) it is more like "cargo cult testing".
Surely but that isn´t different, as you in any case carry a lot of assumptions (like, it´s slightly more likely to get wrong results in "sighted listening",or about the relevance/degree of a difference and it´s associated outcome in experiments, or about the objectivity of experimenters and so on and on....)
Of course it can be.....
As a scientist you already know that it doesn´t invalidate the critic; it is either justified or it is not. You might not like it due to the reasons you´ve mentioned, but that is a totally different matter.
Again as a scientist you already know that it could happen due to good reasons. For example i don´t like the ABX protocol myself, and as written before did notice that two people i tried it with, experienced quite severe difficulties, so i dropped it from my method list. But i nevetheless have to accept that other people might have very different experiences,so if listeners like for example Paul Frindle or Bruno Putzeys report results from ABX tests i don´t have generally a problem with acceptance despite my general concerns.
As a scientist you are able to make a distinction between "bashing" and justified critic based on scientifc evidence. I´m sure we can agree that consideration of the known pitfalls is quite rare despite the fact that these pitfalls are known for at least 60 years.
Of course, ABX is just a protocol, but as there is constant demand for "doing an ABX" now justified by the anticheating part, it is imo the mainly used protocol in the amateur test department.
Which is a completely different topic, isn´t it?
We all learn in our lifes to rate the subjective advice of other people differently according to our experiences with this kind of advice.
As someone who worked in 2 scientific institutes as a scientist during my professional career I know that scientists do not differ much from "ordinary people", many of them have their beliefs and biases and in order to protect their methods some of them are willing to hide some results that do not confirm their theories. Of course not everyone behaves this way, but the scale is as wide as with other people.
- Status
- Not open for further replies.