What people should come to understand is: Audio design professionals like Charles Hansen and I have chosen to make a living designing audio equipment. If we are NOT successful in our design approaches, we will lose our reputations for making quality audio products and either not be hired for new designs or we will go broke, not being able to sell our existing designs.
In my case, I have found it most 'useful' to learn from my past experience, what works and what doesn't work, and rely on the opinions of others, often audio reviewers as to my 'success' with an individual design. Often, a specific design needs improvement to make it all that it can be. On my 'drawing board' right now are potential changes to the JC-1, JC-2, and JC-3, my present Parasound 'flagships'. I want them to work sonically as well as possible, and I find that continual 'refinement' is usually possible and perhaps practical, once I get some 'accurate' feedback from well meaning, but distant sources.
I find that double blind testing is meaningless in regard to my design approach. It just does NOT add anything and it misses a lot. So I will not use it.
In my case, I have found it most 'useful' to learn from my past experience, what works and what doesn't work, and rely on the opinions of others, often audio reviewers as to my 'success' with an individual design. Often, a specific design needs improvement to make it all that it can be. On my 'drawing board' right now are potential changes to the JC-1, JC-2, and JC-3, my present Parasound 'flagships'. I want them to work sonically as well as possible, and I find that continual 'refinement' is usually possible and perhaps practical, once I get some 'accurate' feedback from well meaning, but distant sources.
I find that double blind testing is meaningless in regard to my design approach. It just does NOT add anything and it misses a lot. So I will not use it.
Just lazy, I got a pile of JFET's at Akihabara and built a phono front end. Open-loop linearity was certainly overkill for LP's.
I am designing with 2 x 2SK170BL now. Curious to now results of comparison listening tests vs. complementary differential. Single differential is compromised in a slew rate and its symmetry.
Sorry, you're wrong about that.
I was working with the author of Decibel:
Decibel from sbooth.org
During the beta testing phase he made a change that I said affected the sound. It involved the conversion of 16-bit Redbook audio files from integer to double-precision (64 bit) floating point and back. He said that there was no way for it to affect the sound because the output bits were identical.
So I told him to send me three test pairs and that I would identify them. I got them all right. Of course everyone will tell me that it was just a series of lucky guesses. The chances are 1 in 8. But of course that's never good enough. Nothing is ever good enough.
Enjoy your mid-fi.
Yes, I agree with Charles Hansen, a noted competitor. It gets tiresome to be put down by people who INSIST that I don't know what I am listening to.
I had a similar experience in Japan in 1978. I was asked to listen to 3 separate audio gain stages, NOT designed by me. I picked them out and put them in order. Just guesswork? The Japanese were amazed! It's no real secret, just get reasonably educated, get a degree in physics for example, but don't sell out to the academic establishment. Then get a lot of experience listening to different things, and also listen to the opinions of others, who might teach you something that you overlooked. Maybe then you will qualify as a successful audio designer. '-)
I had a similar experience in Japan in 1978. I was asked to listen to 3 separate audio gain stages, NOT designed by me. I picked them out and put them in order. Just guesswork? The Japanese were amazed! It's no real secret, just get reasonably educated, get a degree in physics for example, but don't sell out to the academic establishment. Then get a lot of experience listening to different things, and also listen to the opinions of others, who might teach you something that you overlooked. Maybe then you will qualify as a successful audio designer. '-)
You do love your extreme examples. I notice he still is using 64 bit processing, I guess he found the trick to make two identical files be identical. If the files were bit perfect copies of eachother how did he tell them apart, Norman coordinate?
Last edited:
I dunno. Could you send me copies of those paired files? You don't have to tell me what your guesses were.
It's amazing, just the same thing is happening at the moment on a german forum, there the claim was that two bitidentical .WAV-files -- rips from a cd -- did (or at least could) sound different when played back on a portable media player, because (and that's the point) one was ripped on a notebook with AC power supplied while the other was ripped minutes later in the same session, but laptop running off its battery.
Currently bulk copies (10 of each rip, ramdomly sequenzed, hidden to the testers of course) of those bitidentical files (I personally checked this) are moved around the internet and then downloaded to the player again to see if the claimed correlation "battery rip"=="sounds better" holds. Aha!
Sort of the hard way to get a proper shave by Bill 'Razor' Ockham himself...
Even audio magazines have picked up similar "themes", like the sound of different ripping software (this time files may be only bitidentical in the part they share in the audio stream because of different start offsets -- a CD drive issue). Still rips from EAC would sound different than the same rips as extracted by foobar2000, they write.
I wonder where this audio world is going to....
Equality in the audio-stream of files is a brick-wall dead end for any event in the history of the files. If any audible difference occurs with identical audio streams played back, it can only come from the actual playback mechanism. Position of the file in the mass storage's file system or RAM is different, this could lead to constantly changing electrical conditions (noisy voltages, jitter) at the DAC-chip, and it is thinkable that a stable correlation between this and percieved sound could take place, the simpler the player's firmware and hardware, the more likely. While this is "far out" as they say, for me at least, it is possible. But bitidentical files "remembering" their history and acting like monsters? This not possible at any rate. If there is need to argue about that point the world would start to fall apart completely....
Currently bulk copies (10 of each rip, ramdomly sequenzed, hidden to the testers of course) of those bitidentical files (I personally checked this) are moved around the internet and then downloaded to the player again to see if the claimed correlation "battery rip"=="sounds better" holds. Aha!
Sort of the hard way to get a proper shave by Bill 'Razor' Ockham himself...
Even audio magazines have picked up similar "themes", like the sound of different ripping software (this time files may be only bitidentical in the part they share in the audio stream because of different start offsets -- a CD drive issue). Still rips from EAC would sound different than the same rips as extracted by foobar2000, they write.
I wonder where this audio world is going to....
Equality in the audio-stream of files is a brick-wall dead end for any event in the history of the files. If any audible difference occurs with identical audio streams played back, it can only come from the actual playback mechanism. Position of the file in the mass storage's file system or RAM is different, this could lead to constantly changing electrical conditions (noisy voltages, jitter) at the DAC-chip, and it is thinkable that a stable correlation between this and percieved sound could take place, the simpler the player's firmware and hardware, the more likely. While this is "far out" as they say, for me at least, it is possible. But bitidentical files "remembering" their history and acting like monsters? This not possible at any rate. If there is need to argue about that point the world would start to fall apart completely....
Jacques Benveniste comes to mind. Nonetheless, I'd love to see Charles's files.
Yes - why would he do that for any reason other than marketing? I can see the value if he's running the audio through SRC or EQ stages, but surely both of those are anathema? Engineering wise it would be far preferable to stick with integer processing - lower CPU overhead, lower power, less EMI. He claims that disk playback results in glitching, perhaps its because he's running out of CPU cycles?
What is sad, is that no matter what Charles and I do, certain people don't believe us. It just comes down to pushing against one person or another's belief system. Or.
"He or she is constantly trying to prove that much of the daily experience of the rest of humanity is "delusion," "hallucination," "group hallucination," "mass hallucination," "mere coincidence," sheer coincidence," or 'sloppy research" " Robert Anton Wilson 'PR'
This seems to happen here on a daily basis. But why? What is the harm, even if Charles and I were wrong? People here gasp at spending $100, certainly our multi thousand dollar class A designs are not on your list.
"He or she is constantly trying to prove that much of the daily experience of the rest of humanity is "delusion," "hallucination," "group hallucination," "mass hallucination," "mere coincidence," sheer coincidence," or 'sloppy research" " Robert Anton Wilson 'PR'
This seems to happen here on a daily basis. But why? What is the harm, even if Charles and I were wrong? People here gasp at spending $100, certainly our multi thousand dollar class A designs are not on your list.
First principles of information theory and by implication the second law. Charles and JC build beautiful boxes with impecable power supplies, low noise, and circuits as devoid of artifacts as is humanly possible. Why they need to harp on these eccentric contrarian beliefs is beyond me. We don't understand everything so everything we know is wrong, I don't buy it.
I'm much more interested in finding out that one of these profoundly odd beliefs are right. "There's a pony in here somewhere."
I have a little trouble with the idea of bit identical large files. If you do a true bit by bit comparison of commercial CD's there are always differences in the errors. (Last I talked to someone in manufacturing they were using optical computing to measure the errors at production speeds and the acceptable pinhole and other error rate was well above zero.)
Even in more robust media there is still an error rate. So just as it is hard to believe anyone could pick out a single LSB error in a large file it is just as hard to believe you have a truly error free large file and two copies are identical. Yes I know about parity and ECC so it is possible to create such files, but I am unaware of anyone going to those lengths.
Even in more robust media there is still an error rate. So just as it is hard to believe anyone could pick out a single LSB error in a large file it is just as hard to believe you have a truly error free large file and two copies are identical. Yes I know about parity and ECC so it is possible to create such files, but I am unaware of anyone going to those lengths.
- Status
- Not open for further replies.