I am lucky John, I got my education in some different schools, so my expectations were different: we were learned to use math, physics, and own gray mass to solve the problem, instead of books with "all right answers" and patent libraries.
Not quite the same ... but worthy of mention. The Armstrong feedback circuit filed 31/01/13 was an excellent example of feedback being disputed in patent courts and eventually decided in Armstrong's favor, despite Supreme Court rulings backed by Lee De Forrest and backed by AT/T. Edwin Howard Armstrong - Wikipedia, the free encyclopedia.
and: Living Legacies
Cheers / Chris
Attachments
Ouch.
Hey John, up your neck of the woods. We talk about physical layout and other issues. Last time I was up at LBL, the light source was down for calibration. We got a tour inside the ring. What an incredible mass of unlabled cables, crossed up, twisted and miles of aluminum foil trying to shield or insulate everything. It must take a dozen grad students per experiment just to know what is what. Yet world class results come out of there every day. What is pretty scary is seeing the 6 inch thick lead ring there just in case of a magnet failure as the beam would cut the foot of concrete right in half. And the building, and anything else in a few miles.
Hey John, up your neck of the woods. We talk about physical layout and other issues. Last time I was up at LBL, the light source was down for calibration. We got a tour inside the ring. What an incredible mass of unlabled cables, crossed up, twisted and miles of aluminum foil trying to shield or insulate everything. It must take a dozen grad students per experiment just to know what is what. Yet world class results come out of there every day. What is pretty scary is seeing the 6 inch thick lead ring there just in case of a magnet failure as the beam would cut the foot of concrete right in half. And the building, and anything else in a few miles.
Here, I give you a quote, without comment:
"In whatever way we experiment, we obtain results which are somewhat as follows:
The timbre depends only on the relative energies of the various harmonics and not on their phase-differences. Differences of phase produce no effect on the ear. This is known as Ohm's law, having been discovered by G.S. Holm (1787-1854), the discoverer of the still better known electrical law." 'Science & Music Sir James Jeans 1937.
See many of you are in 'good company'. Why do we not give OHM his credit for this discovery?
"In whatever way we experiment, we obtain results which are somewhat as follows:
The timbre depends only on the relative energies of the various harmonics and not on their phase-differences. Differences of phase produce no effect on the ear. This is known as Ohm's law, having been discovered by G.S. Holm (1787-1854), the discoverer of the still better known electrical law." 'Science & Music Sir James Jeans 1937.
See many of you are in 'good company'. Why do we not give OHM his credit for this discovery?
Could be because the very best science of 1842 may not be the very best science in 2012. Could be correct though, as if I remember, Georg Ohm was a pretty bright fellow. Now, he and Helmholtz stated that only constituent frequencies and their amplitudes produced our perception of a tone. They did not make any assertions on the relevance to phase in localization.
We can localize a pure tone in the horizontal plane.
We can not localize a pure tone in the vertical plane. ( in a chamber with no reflection effects. )
We can localize a complex signal in mono played between two speakers, and you can play with just amplitude and make the image move between the speakers. Cover one ear, it does not move. Neat trick. People with only one good ear can still localize sound somewhat, just as you still have some depth perception with one eye.
Does phase matter? I heard a Grover Washington record played through a top line Polk speaker with some of his phase/image whatever he was doing. It sounded like he was running fore and aft on the stage. Bout' fell out of the chair laughing. If phase does not matter, why can I hear localization of sounds outside the width of my speakers.
My message: what and how our brain processes as sound is still barely understood by the best of science and even less by me.
We can localize a pure tone in the horizontal plane.
We can not localize a pure tone in the vertical plane. ( in a chamber with no reflection effects. )
We can localize a complex signal in mono played between two speakers, and you can play with just amplitude and make the image move between the speakers. Cover one ear, it does not move. Neat trick. People with only one good ear can still localize sound somewhat, just as you still have some depth perception with one eye.
Does phase matter? I heard a Grover Washington record played through a top line Polk speaker with some of his phase/image whatever he was doing. It sounded like he was running fore and aft on the stage. Bout' fell out of the chair laughing. If phase does not matter, why can I hear localization of sounds outside the width of my speakers.
My message: what and how our brain processes as sound is still barely understood by the best of science and even less by me.
But of course phase is not a useful concept when talking about transients and non-harmonic distortion - which seemed to be the area of amplifier performance that people were homing in on a few pages back. Why this urge to express everything in the frequency domain?
When we calculate THD, or THD+N, we are, effectively, converting a frequency domain measurement back into an absolute measurement (sum) of deviation from perfection in the time domain. The frequency domain is just a useful tool in thinking about things, and a very sensitive way to reveal distortion even if all we have is analogue measurement gear. But if the difference between amplifiers with identical non-transient measurements is in their handling of certain (possibly arbitrary) transients, the frequency domain may not be the best way to think about it.
When we calculate THD, or THD+N, we are, effectively, converting a frequency domain measurement back into an absolute measurement (sum) of deviation from perfection in the time domain. The frequency domain is just a useful tool in thinking about things, and a very sensitive way to reveal distortion even if all we have is analogue measurement gear. But if the difference between amplifiers with identical non-transient measurements is in their handling of certain (possibly arbitrary) transients, the frequency domain may not be the best way to think about it.
A thought experiment: could you devise an amplifier circuit that passed all the usual tests (low distortion, flat frequency response etc.) but sounded terrible? My initial thought is an amplifier with some sort of in-built dynamic range compressor. What tests could guarantee to reveal the presence of the offending circuitry?
If you restrict your tests to small signal then one could name dozens of things that could do it, but compression makes distortion and unless your amplifier topology is deliberately designed to fool it a high level IMD test would not pass.
Last edited:
AFAIR Helmholtz was misinterpreted (to a certain degree) as the conclusion of ear phase deafness was based on his work. Furthermore even at that time there did exist a diverging theory, but for the next ~140 years the ´phase deafness´of the human ear got somehow accepted as a fact.
Experiments with the polarity of signals make some people wonder about this "fact" and since 1998-1999 it is known that the human ear is not phase deaf .
Quite consistent on both sides, i.e. listening tests and studies regarding ear physiology.
Experiments with the polarity of signals make some people wonder about this "fact" and since 1998-1999 it is known that the human ear is not phase deaf .
Quite consistent on both sides, i.e. listening tests and studies regarding ear physiology.
Funny thing, because during the ´70s and ´80s the number race was on and a lot of (especially japanese) amplifiers reached extremely low THD+N numbers while maintaining of course flat frequency response.
It is a matter of combining these amplifiers with loudspeakers and "sounds terrible" is of course not really true, but there were a lot of combinations not delivering a convincing performance.
A couple of weeks ago (wasn´t it the Blowtorch thread?) the Quad amplifiers were mentioned; at that time it delivered good measurement results (at least compared to the known hearing thresholds), but wasn´t that good for careful listeners. For example to my ears it was not the best (not even a sufficient) choice for Quads own electrostatics.
Today it is argued that modern high frequency distorsion tests (duo tone 19kHz + 20kHz ) can reveal the flaws of the amp, which is obviously true, but back in the ´80s it was considered to be a decent amp.
So, basically the arguments were always the same, but looking back after 20-30 years it seems that some of the "golden ear" arguments might have been correct.
Last edited:
Some people seem to be mixing up transient and thermal effects. Transients are fast - the issues are things like slew rate. Thermal effects would be slow - LF and signal envelope.
For the avoidance of doubt, I am not claiming that phase does not matter. It clearly affects the stereo effect, and may have some role in other areas too. All I said was that the presence or absence of frequency components is the main issue, as Mr Ohm found all those years ago. Accordingly, I believe that a weighted THD is a useful number; unweighted THD less so. If we knew more about masking we could probably come up with a better weighting system. The main benefit of THD-type numbers is that they also give a guide to the likely level of IMD, as both are caused by non-linearity. IMD is worse than harmonic distortion because it is not harmonic.
For the avoidance of doubt, I am not claiming that phase does not matter. It clearly affects the stereo effect, and may have some role in other areas too. All I said was that the presence or absence of frequency components is the main issue, as Mr Ohm found all those years ago. Accordingly, I believe that a weighted THD is a useful number; unweighted THD less so. If we knew more about masking we could probably come up with a better weighting system. The main benefit of THD-type numbers is that they also give a guide to the likely level of IMD, as both are caused by non-linearity. IMD is worse than harmonic distortion because it is not harmonic.
Was the difference blindingly obvious, or subtle? Remind me (it's a while since I read the relevant thread), could you tell which was straight and which was processed?
An all-pass will cause a delay in the frequency region where the phase is varying so it could be the delay you hear rather than the phase.
An all-pass will cause a delay in the frequency region where the phase is varying so it could be the delay you hear rather than the phase.
I once had an idea, to simulate pipe bass, adding inter-modulated APF for bass guitar, kind of Doppler effect simulation. The result was horrible! More like listening SSB modulation on AM receiver than Doppler effect. After that experiment I learned to hear phase intermodulation, even if it is subtle.
I don't believe in simple standard tests to measure all possible errors of all possible topologies. Even marketing departments don't care today to show results of simple tests, they invent something new to catch up imagination of potential buyers, like synesthesia - based epithets.
- Status
- Not open for further replies.