But of course, results are valid in those test conditions. That is always the case.
We take care of that in the test protocol. The whole *point* of controlled tests is to eliminate the witness subjectiveness.
Tech progress has brought you audio technology, no question about it. What else would have?
I don't think SY claims his tube amps, while being clever designs and, in that niche certainly discover better/more effective/higher quality equipment, furthers 'audio technology' per se.
The fact that many are content with coloring, low efficiency, high distortion tube se amps doesn't mean that audio tech as such hasn't certainly progressed a LOT beyond that.
jan didden
. Results like these are no use for guiding the design of NEW amplifiers - which must present new Test Conditions! Making a judgement on new caps in old amp designs is a very limited scope, and limited interest.
Eliminate witness subjectivity? You have a protocol that makes all witnesses have the same taste? Can it make them all want to marry the same woman, also?
My point is that the progress has been doubtful, especially since the 1980s. Even today, buying a new recording (CD) of a long Opera is an unrewarding experience, and often I can't listen all the way to the finale, because the sound is so fatiguing. This is not progress in my view, considering the quality of 50-year old LP operas, just for one example. The measurements of THD, or whatever else we have today are seemingly not enough, and I say this as a professional electronics designer, not a dilletante.
It is pretty difficult to perform "controlled" tests.
Regardless switchboxes, that may influence audio signal (EMI/RFI content etc.), it is difficult to assure good listening conditions. Imagine set of 10-20 people, sitting somewhere, where it is far from optimum sweet spot. There is usually a very limited space where we can get a point with perfect spatial information. IMO these "controlled" tests are as much subjective as any other subjective sound description provided here. As Rod has already said, they are valid only for a specific place, acoustics, instruments - specific test conditions. The result is not transferable. So - is it any "controlled" test, is it any "scientific" method?
Regardless switchboxes, that may influence audio signal (EMI/RFI content etc.), it is difficult to assure good listening conditions. Imagine set of 10-20 people, sitting somewhere, where it is far from optimum sweet spot. There is usually a very limited space where we can get a point with perfect spatial information. IMO these "controlled" tests are as much subjective as any other subjective sound description provided here. As Rod has already said, they are valid only for a specific place, acoustics, instruments - specific test conditions. The result is not transferable. So - is it any "controlled" test, is it any "scientific" method?
And I was just talking about measurements that showed an electrolytic capacitor used as a DC blocking capacitor in the feedback path shows distortion of .06% THD versus .005% for a film capacitor in the exact same amplifier at the same level and load.
Can people hear .06% THD (-65db), If you say you can or can't I see no reason to disbelieve you. Of course good engineering says to use the lower distortion part unless the $.37 versus the $3.83 is important. Then you actually have a decision to make!
One of those important lessons I learned in College, but not in a class, was when you set up the mechanism to do the work easily, it gets done more often.
Here it is easier to speculate and complain, than it is for most folks to actually do experiments.
So which has more value?
I'll take a badly done but well meaning test over an uniformed opinion almost any day. (Depends on what she looks like!)
Can people hear .06% THD (-65db), If you say you can or can't I see no reason to disbelieve you. Of course good engineering says to use the lower distortion part unless the $.37 versus the $3.83 is important. Then you actually have a decision to make!
One of those important lessons I learned in College, but not in a class, was when you set up the mechanism to do the work easily, it gets done more often.
Here it is easier to speculate and complain, than it is for most folks to actually do experiments.
So which has more value?
I'll take a badly done but well meaning test over an uniformed opinion almost any day. (Depends on what she looks like!)
Last edited:
.
Eliminate witness subjectivity? You have a protocol that makes all witnesses have the same taste? Can it make them all want to marry the same woman, also?
There's difference and preference, on teflon vs PP caps we have not gotten to difference yet (which does not involve taste).
Well said.
Since I don't design amplifiers, I'm left with choosing between available ones and trying some DIY.
I believe that most people choose loudspeakers by listening to them, not by reading their datasheet. With speakers it's quite obvious and widely agreed that their data doesn't say much about how they sound. But when it comes to amplifiers, many people believe that their data is all there is to know about how they sound, which is a false belief. In his respect, amplifiers aren't different from loudspeakers.
With all audio equipment, be it loudspeakers, amplifiers and sources, the choosing has much to do with personal taste and preferences.
Differences are also interpreted by ears and Brains. Do you believe that all listeners will identify all the differences? And report them as being the same? What you say highlights even greater difficulties in attempting to use scientific methods to tackle problems outside of their realm.
Hmmm, for those of you who want a basis for some standardised tests, try the iec.ch website, they used to have a lot of the old standards available for free, from memory some of the P.800 series have detailed information about the listening room, listening conditions and durations allowed. BUT I tried looking just now at the website and couldnt find them.
Wrinkle
Wrinkle
I did say audio measurements? You could easily set up a chopper front end and do a DC to a kHz or so. I thought this might be good for ULF radio.
This problem is not outside the realm of science. I'm saying that in some cases no listeners will tell any difference, such as Greiners null result on filtering past 20kHz.
- Status
- Not open for further replies.