You are scratcing the core of a long standing –still hot- debate among the supporters and the opponents of Plato’s expressed ideas over “knowledge”. Don’t expect to reach a consensus in this lifetime
(unless you know that already ) George
Well, now I'm thoroughly confused. You published them, you didn't publish them, now you published them. When you figure out if and where where you've published detailed descriptions of test protocols, methods, controls, data, results, and analysis, I'll be interested to read them.
At that time i read the comments of the anonymous professor in the Audio Critic, but had not read (and have still not) Hawksford´s article and afair the anonymous got at least two things wrong in his rather short statement.
The funny thing is, that according to jneutrons arguments, the anonymous professor might have missed a lot more.
So it´s time to read the original papers.
My take on the prof is that he wrote his comments while emotionally charged. As such, I am guessing that he didn't feel the need to elaborate past a certain point due to the obvious nature of the errors. I am also guessing that the prof was exceedingly unhappy that such an article was printed within the popular press without the benefit of peer review. Peer review may have turned up the failure to consider internal inductance that I discuss.
To me, neglecting the internal inductance is a clear and obvious test violation, but since I live and breath this kind of test stuff, my experience may be the only reason I spotted it.
When I first encountered his published waveform, I was shocked that it was an exact duplicate of one I dealt with a few years prior. The upshot of my analysis and fix was that the diode being tested did not have a Trr of 5 milliseconds, but that the 250 micro-ohm CVR was collapsing it's field and the test system was picking up the flux of the CVR. Since the CVR was a set of 11 copper plates 4 inches by 4 inches by about 100 mils, I simply had the tech run a wire down the geometric center of the resistor...that eliminated the flux trapping and solved the problem. (pic attached, it's fuzzy cause it's through a 1/4 inch lexan blast shield..)
It did not require I define some new entity via planar wave field penetration to describe the "new phenomena", just an understanding of ampere's and faraday's law and how they can bite ya if you're not careful.
Cheers, jn
ps. The article initially cited was published in stereophile october 1995, starting on page 53, ending on page 69. The 70 page document mentioned by John Curl was called a Unification document, comprised of many "tracks", and on my computer it comes out as 88 pages. Page 70 is where the conclusions begin, at the bottom...and the following pages are appendicies, test diagrams, and matlab code he generated for his simulations. The unification doc was a series of publications in HFN/RR, starting in may of 86.
Attachments
Last edited:
People always seem to be surprised, and cry "distortion!", when the waveform at the output of a filter is different from the (non-sinusoidal) waveform at the input. Did they expect just a change in amplitude and a bit of time delay? Hawksford carefully navigated around this particular mistake, by admitting that the system was linear and so could not generate new frequencies, but then threw a bone to the cable freaks by muttering about grain boundaries etc. Trying to have his cake and eat it?
Maybe someone on here with the necessary software to hand could publish the spectrum of a short unprocessed zero-starting (and finishing) tone burst, so that others can see what a wide band it covers. To make life easy, you could assume a burst repeat frequency which is an integer sub-multiple of the tone frequency. Then put it through a first order low pass filter with rolloff well above the tone frequency, to exhibit the 'distortion and time smearing' which appears at the leading and trailing edges. I think this would be educational for those who have never actually calculated a Fourier series themselves.
Maybe someone on here with the necessary software to hand could publish the spectrum of a short unprocessed zero-starting (and finishing) tone burst, so that others can see what a wide band it covers. To make life easy, you could assume a burst repeat frequency which is an integer sub-multiple of the tone frequency. Then put it through a first order low pass filter with rolloff well above the tone frequency, to exhibit the 'distortion and time smearing' which appears at the leading and trailing edges. I think this would be educational for those who have never actually calculated a Fourier series themselves.
I scored 100% without even listening to the music at an AES demonstration by listening for the A to B and B to A clicks.
Do you think folks will ever stop trying to PROVE digital audio is flawed by using waveforms that violate the sampling theorem?
Let me guess he sited virtually none of the vast amount of literature on the matter. Van der Ziel is required for starters.
For amusement - http://en.wikipedia.org/wiki/Quantum_1/f_noise The section on the controversy sounds familiar
Last edited:
I think this is consistant with what I've been trying to say, just not nearly as concisely as you've put it here. Well said.
Overhere, there's a saying that something tastes as if an angel is peeing on your tongue.
Clearly, Heineken hired someone else to do the honors.
(Evidently, there's a need for a stupiderest reference posting on this thread)
Sorry I was wrong none at all, stunning nonsense here too. JN please point out the moments of brilliance.
Yes, I think that is what he called it. It had me fooled for a while, but then I started thinking about the consequences if it were true. As I said, partition noise at a circuit junction? For this reason, I now take everything he says with a big pinch of salt. I don't believe he sets out to confuse people, but I suspect his grasp of physics and maths is not as reliable as he believes. He also seems to enjoy playing the eccentric maverick, so criticism has little effect. Essex University always had a bit of a reputation for being radical, outside the mainstream.Chris Hornbeck said:
You have me over a barrel here. Obviously I did not mean any of his wire work nor capacitor work.
I remember reading some of his papers, I think either digital or feedback or current drive
, and remember coming away very impressed. For the topics, the papers seemed internally consistent, and didn't counter what knowledge I had previously understood. So in that respect, I did think the work very good, brilliant as a descriptor seemed to fit.For papers such as the noise one being discussed here, I am not in a position to understand the ups and downs, needing to take a back seat to those well versed.
I would NOT like to think that all of his professional work fell apart like a cheap suit under scrutiny the way his early essex echo article does. But I'm afraid that I couldn't ascertain that for topics way over my head. I'll try to find one of the papers I was impressed with..
Cheers, jn
That's OK, I also found value when he stuck to circuits, etc. Back of the envelope, -100dB on a 1mV signal at midband from a MM cartridge is still around 10^7 electrons per second.
EDIT - It appears MH's argument collapses when you consider a JFET input .
Last edited:
The number sounds about right. But a second is a fairly long time in audio; at 10KHz that's 10^3 electrons per cycle, from the generator. How many get injected will vary with Beta, yada yada. Not an issue with field effect devices like electron valves or those new-fangled whatchacallits.
Thanks,
Chris
But this is only the signal which is added to the random thermal noise. I think I estimated the noise for a shorted 10 Ohm MC at about 4nA rms 20-20K this is a LOT of electrons. It is MHO that these arguments reduce to treating the electron solely as a particle which leads to wrong conclusions.
Well if MOND theory pans out we don't need the dark matter. Newton and Ohm work well enough for me.
- Status
- Not open for further replies.