Thanks, SY, I've started to delve more seriously into the material. First of all, some mild amusement: in spite of all that has been said here, from Audio Critic 20, talking of assessing power amplifiers ...
Moving on, an interesting new term, to me, "dynamic distortion". Once I understood more eactly what was being referred to, it turned out to be a very old friend, in a different guise - this is something I'm very sensitive to, why I find most playback lacking at the first hurdle.
Moving on, an interesting new term, to me, "dynamic distortion". Once I understood more eactly what was being referred to, it turned out to be a very old friend, in a different guise - this is something I'm very sensitive to, why I find most playback lacking at the first hurdle.
Possibly, perhaps probably. My aim was to avoid crossover distortion.fas42 said:
I assumed that the distortion would be dominated by 2nd and/or 3rd order. Perhaps I should have said that explicitly. In almost all circuits this will be the case.
I was trying to express in the form of a performance figure the type of result you would get from an amplifier whose in-out characteristic is polynomial or convergent infinite series, with large low order coefficients and small high order coefficients, with strict monotonicity and infinitely differentiable. That is, smooth but not necessarily completely straight.
Outputs compared with inputs were done by Quad many years ago as an advertising stunt. They wired 10 (or 20?) power amps in cascade with attenuators, then showed (at a HiFi show?) that apart from a little noise the sound was indistinguishable from that from one power amp. This means that one of the following is true:
1. their amp does not noticeably distort the sound, even when used many times.
2. their amp distorts the sound on the first pass but subsequent passes do not make it any different (it is difficult to imagine how this might be achieved).
An example how to test:
In this case its an AD/DA converter, but you can do the same with any device.
As DF already has mentioned, there are lots of tests done to come up with some numbers, but they are a bit fuzzy. Engineer well below them, not that difficult, and its very likely (but not 100% certain!) your safe. Listening tests are always required to make sure.
This not a test. How are you going to give a value to the results? It sounds chunky, meaty or fresh and to what a degree?
This is an ABX setup, therefore it can only test for audible differences and nothing more. You give value to the test by repetition and statistical analysis.
Chunky, meaty or fresh are not exactly defined and therefore useless for testing.
Let's try again ...
suppose I design and build what I believe is a "faultless" amplifier, and I wish to convince an appropriate member of the scientific community of that fact, merely by sending him a set of measurements and graphs, what data would he have to see, as a minimum, and what would the numbers be - the sufficient case
Sorry Frank, a given amplifier's performance only makes total sense in situ.
Exactly.
For explaining audible differences objective measurements can be used.
For giving audible differences a value other subjective testing methods are used like ABC/HR or MURSHA (and others I don't know of). These are much more elaborate.
That's the idea. Before trying to explain an audible difference or try to correlate a measurement (or decide the appropriate measurement), establish that one actually exists.
We are doing the test on a Monday morning after a weekend spent at a rock festival, or are we doing the test on a Wednesday evening with all the lights turned off and somebody has educated us on what to listen for.
The results of such a test would be: Yes I can hear a difference or no I can not hear a difference and the results would tell you more about the person doing the test than the equipment being tested.
If you tested a large number of designers/listeners you could select the best ones to do the listening test. It would give you some kind of reliability to the testing.
The results of such a test would be: Yes I can hear a difference or no I can not hear a difference and the results would tell you more about the person doing the test than the equipment being tested.
If you tested a large number of designers/listeners you could select the best ones to do the listening test. It would give you some kind of reliability to the testing.
In you mean in the context of a complete system, then I totally agree. Therefore, the audio community, yet again, has only itself to blame for all the silliness that abounds, by not making it absolutely clear what all the requirements are for getting satisfactory sound. Just saying something like "Use an adequate amplifier" is equivalent to "Find a piece of string that someone says is long enough".
I totally support the concept that only a complete system's performance is meaningful, at the moment - simple measuring of one aspect or component of it, in a way that suits that device, is only a small part of assessing competence.
That was a "stunt" but they were heading in the right direction: the fundamental flaw in the exercise was that each amplifier was not heavily loaded at its output, forcing the unit to deliver significant current. Of course, the end result would have sounded terrible, not much use as an advertising gimmick ...
!!But, a competent amplifier should be able to pass such a severe loading scenario, in a listening test - otherwise it is faulty, end of story ...
within the bounds of reason "adequate" is probably the only qualifier you can use if your aim is "satisfactory" sound.
I am reminded of the possibly apocryphal story about the power ratings Bentley gave for their automobiles. I'm told that the Bentley manuals list power as being "adequate" except in the case of the Mulsanne which is described as "more than adequate".
I am reminded of the possibly apocryphal story about the power ratings Bentley gave for their automobiles. I'm told that the Bentley manuals list power as being "adequate" except in the case of the Mulsanne which is described as "more than adequate".
If that is how listeners hear it then they are not listening to the right things; in the same way people can learn to listen for codec compression artifacts, and easily pick misbehaviour, so can one learn to discern conventional audio playback misbehaviour - what is typically called the "sound" or signature of components. The real point of the exercise then becomes to determine which setup is least interfering with the source waveform, in the sense of how the final sound registers subjectively.
- Status
- Not open for further replies.