|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
83.80.191.138
In Reply to: RE: I don't know why you didn't find my posts posted by Tony Lauck on June 14, 2010 at 15:08:58
So, that's settled then. Which means that in fact the setups used for the test were ok (source material (check out David Moran's posts in the linked thread), gear, method), except they did not meet "audiophile standards". Unless I missed something, nobody pointed towards a real flaw AND PROVIDED SOLID EVIDENCE. Just assumption and opinions. As for allegedly insufficient resolution of gear, room, listener, where does it say that you need Harry Pearson's personal system?
However, as far as the listener is concerned, I agree with J. Gordon Holt:
"The listener is the heart of the high-fidelity system, and is noted for having high distortion, poor frequency response, marginal stability, and arbitrarily variable performance characteristics. Listener instability is the most common form of defect in a high-fidelity system, which is why manufacturers recommend that the ears be checked periodically by a qualified service agency to ensure that they meet their specifications. Defective ears may be cleaned with anti static spray or a mild washday detergent containing a wetting agent, or may be replaced by a microphone and an oscilloscope or, in cases where there is little interest in music, by a camera and a well-equipped dark-room."
Klaus
Follow Ups:
The real flaw is that there was no calibration of the source material, equipment, room, etc. as to resolution and suitability to detect the effects being evaluated, nor was there qualification or training of the listeners. For these reasons the failure of the tests to produce results has little predictive value as there is no reason to extrapolate the results to other conditions and subjects. Worse is the interpretation that others have placed on the null results. (Meyers and Moran are somewhat circumspect in this respect.) The failure to disprove the null hypothesis is just a failure to detect. One can not conclude anything from this statistical evidence.
You've got the burden all mixed up. The authors showed nothing, so there is nothing to rebut. What is curious is that the JAES even bothered to publish this non work.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
1. Calibration: If we call the effect to be evaluated "bottlenecking", the possibly only acceptable way to calibrate the system is to use a bottlenecked signal, hence the effect itself, otherwise you probably would say that the signal used for calibration was not suited. If you know of any other test material with known degradations equivalent to bottlenecking other than bottlenecking itself, go ahead. The next question is, how do you determine those “known degradations”? By biased sighted testing, by non-calibrated blind testing?
2. Qualification and training of listeners: no audiophile is calibrating his system, no audiophile is trained, are you willing to accept all those positive results of all those heavily biased listening tests as conclusive evidence? Brad's test included recording engineers and a mastering facility, and audio engineering students and a custom-designed university audio facility, if that's not trained listeners and room acoustics with sufficient resolution, I don't know. They also included an audiophile with his audiophile system and custom-built listening room, at least that should do, or is Sea Cliff the minimum requirement?
3. Meyer/Moran being circumspect: they are because it's the only way to look at their results: "our tests failed to substantiate the claimed advantages of high-resolution encoding for two-channel audio" and "it is very difficult to use negative results to prove the inaudibility of any given phenomeon or process".
4. The authors showed nothing: by the same token, nobody shows anything, never. All sighted listening tests are biased. Many audio systems have non-flat amplitude response, on and off-axis. The overwhelming majority of loudspeakers is not time aligned. If you want to look for flaws, look at audiophiles, their systems and listening tests.
5. You've got the burden all mixed up: that's merely a matter of opinion. The study was published in 2007, we are now almost 3 years later, to this date no one has shown, in a test that meets both audiophile and non-audiophile requirements, that redbook is not sufficient/transparent. Instead of poking for holes in the Meyer/Moran test, the audiophile community should do its own test, watertight, foolproof, at Sea Cliff, if necessary and get it published. If you think there's a serious flaw in the study, write a comment on the AES forum, Clark is a member, so there is access!
I for one have said more than I wanted to, so I’m stopping at this point. To me the study is just providing an additional reason for not buying hi-rez.
Klaus
Audiophiles are not publishing papers in "esteemed" journals. What they do or do not do is irrelevant to my criticism of authors or journals.
If one is exploring format limitations, one can very easily "calibrate" the test setup that I previously suggested. (Hi-res PCM vs. Hi-res PCM degraded through a lower resolution "knothole".) For example, one can test the effect of various sampling rates by down sampling and re upsampling to various alternative sample rates. Similarly one can use different bit resolutions. The degradations are all quite precise as they can be software controlled and if one wants one can look at the actual source code that's doing the mangling and vet the algorithms.
I encounter the lack of resolution of RBCD almost every time I work with higher resolution recordings and have to fit 10 pounds of sound into a 5 pound RBCD bag. The result almost never sounds as good as the original. This is true whether the original was reel to reel tape, cassette tape, or high resolution PCM. Just yesterday I spent several hours downsampling a double album that was in 48/24 format to 44/16 so it could be released as a CD. As expected, the final version did not sound as good as the original, despite fiddling with various parameters such as gain, filtering and dither. My views on this subject are hardly unique. (See link)
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: