The most prestigious law school admissions discussion board in the world.
|
 |
 
|
The Penis can you think of one reason for hi res audio more than anti aliasing?
| raped/touched/licked/fucked | 09/21/26 |
| The Penis | 09/21/26 |
| raped/touched/licked/fucked | 09/21/26 |
| The Penis | 09/21/26 |
| raped/touched/licked/fucked | 09/21/26 |
| raped/touched/licked/fucked | 09/22/26 |
| raped/touched/licked/fucked | 09/22/26 |
Poast new message in this thread
Date: September 21st, 2026 8:56 PM Author: raped/touched/licked/fucked (Prestige Feuds) After conceding no one can hear 40 khz (big!), does your whole argument boil down to the filter, reconstruction, and anti aliasing argument as to why we need more data above the range of human hearing? You said the 53-47 test proves UHD is 'better'. Can you explain what exactly makes it better? (http://www.autoadmit.com/thread.php?thread_id=5905832&forum_id=2\u0026mark_id=5310915#50148424) |
Date: September 21st, 2026 9:06 PM Author: The Penis 53/47 doesn’t prove hi-res is “better.” It's evidence that under the conditions that were tested, the two formats were discriminable slightly above chance.
Nobody needs to hear 40k for that to happen. Higher sample rates change the engineering constraints of the conversion and processing chain. Filter transition bands are one example; nonlinear DSP is another, since harmonics generated during processing can alias back into the audible band at lower sample rates. Higher bit depth is another issue and mainly gives lower quantization noise and more headroom during production.
For final consumer playback of the same master through a competent modern DAC, I agree the practical advantage over Red Book is probably tiny or nonexistent for most people. What I disagree with is your claim that any measured audible difference is mathematically impossible.
I'm not even a hardcore audiophile. I used to buy expensive DACs at one point and always thought it was BS, couldn't tell shit in terms of difference. Expensive headphones and speakers definitely make a difference though. For this particular issue I'm not crazy against you, I just think you saying mathematically and physically impossible is stronger than the actual evidence. |
 |
Date: September 21st, 2026 9:20 PM Author: raped/touched/licked/fucked (Prestige Feuds) I appreciate that you acknowledge the case for having more data that we cannot hear or detect beyond looking at a spectrogram and file size in kilobytes and knowing that it's there pertains solely to filtration methods and reconstruction.
You have not at any point mentioned the increased dynamic range of eg 24/96 nor the limitation of Red Book CD at 96 decibels, which is smart because this argument is also incredibly dumb. You've avoided it entirely but it's a cornerstone of what UHD defenders believe. The average dynamic range of a commercial recording is like 10 dB. You would have to go to great lengths to exploit the maximum on a CD let alone UHD which would require capturing sound far louder than the maximum sound pressure level of a studio microphone.
When I say mathematically impossible it is a bit of a misnomer. Of course using scientific tools I'm sure you can find some numbers that differ from the two formats. But as I'm sure you'd well agree, the only test that matters is whether that difference is in any way detectable by human ears. Even if you could prove some marginal benefit in the anti aliasing filter numerically, it would be moot if it was undetectable to humans.
Would you agree that if I were able to prove that any limitations in anti aliasing and filter reconstruction and jitter of Red Book CD were moot and undetectable by humans that at that point it would be functionally impossible for larger bit buckets to produce a 'better' result? Another way to think of it is that if linear DSP and quantization don't cause a problem humans can detect, how could any amount of more data have an effect? In order for hi res to be demonstrably superior, something would have to be wrong with CDs. We may know some things are wrong in the technical sense, but are they wrong to human ears? |
|
Date: September 21st, 2026 9:36 PM Author: The Penis Yes. If you establish that converting the same master to CD quality introduces no audible error under the playback conditions being discussed, I agree that higher resolution offers no audible benefit there. I already accept that conditional. The question is whether you’ve established its premise.
Also I explicitly mentioned higher bit depth and lower quantization noise. Increasing bit depth lowers quantization noise relative to the same max signal. It doesn’t require recording louder sounds or exceeding a microphone’s max sound pressure. Whether that lower noise is audible during playback is another issue.
The meta analysis we talked about in the other thread reported evidence of discrimination under tested conditions. That doesn’t necessarily establish that listeners preferred hi-res, or that consumers need it. I’m open to evidence that the result reflects experimental flaws. Showing that particular filter and jitter effects are inaudible would weaken those proposed explanations; you would still need to address how the experiments produced their results.
Calling “mathematically impossible” a misnomer brings us closer to agreement. My objection was to treating universal inaudibility as already proven. |
|
Date: September 21st, 2026 9:45 PM Author: raped/touched/licked/fucked (Prestige Feuds) I've already given you at least one major hypothesis as to how a listener could differentiate between the two source formats and you never responded. It's like a blindfolded guitar reviewer asked to tell which brand of Strat he's playing just by playing it and by the sound. If his arm brushes the controls he might get a strong clue about the manufacturer, bypassing the sound test entirely. It's like that. There are potential tells. I also think you're overreliant on one study. Is this study repeatable? Can we do it again somewhere else in the world and get 53% every time?
I will find you the information on why all the filtering and anti aliasing stuff is inaudible. It's beneath the sound floor any human can detect. If established it will become silly to hypothesize that the study is right, rather than the fact that UHD is undetectable and there must be some audible flaw in red book CDs. We're both starting from opposing ends of the argument.
Either you have to prove there is an audible distortion in CDs that can be corrected by having a higher bit bucket and special recording tools that capture sound in the inaudible range and have a massive amount of bit depth that can never be used for a practical purpose, or I have to prove that something about a file with more kilobytes than another could potentially be detected in a repeatable study by a very tiny margin for reasons other than audio quality. That's where we're at. |
|
Date: September 22nd, 2026 1:51 AM Author: raped/touched/licked/fucked (Prestige Feuds) In addition to the loading times being a known/obvious "tell" of larger file sizes, another obvious one is controlling for level matching. In virtually any study in which one mix is 1 dB higher than another, respondents will choose the slightly louder mix at a much higher rate. Level matching reduces most people's ability to tell anything in audio apart, even something like a waveform that is uncompressed vs a compressed one people will choose the thing that is slightly louder in a blind test. And it's very hard to accurately control for this. The possibility that the test got absolutely everything right is slim.
Do you think you could tell apart a $100, a $500, a $1000, and a $2000 DAC apart if they were perfectly level matched? I doubt it. Most people could not 'hear' a difference between these components even if some people claim they can. And statistically a 20x jump in price between devices would likely have more of a difference in sound quality than whether or not a source file could reproduce a wave form at 40 khz or whether 24 bits of depth sounded slightly better when quantized. |
|
Date: September 22nd, 2026 2:30 AM Author: raped/touched/licked/fucked (Prestige Feuds) Another key angle here. The guy who did the study Josh Reiss is an academic who founded 4 different audio startups after his research struck various nerves in the British press, including one public feud where he was called out by a member of The Jesus and Mary Chain. Reiss readily admits to being an entrepreneur who uses his research to found audio companies. At the time of his study in 2016 there was a lot riding on the success of Tidal, Pono, and formats like MQA. This does not discredit his work entirely but it's obvious he is very intertwined with the business world and his research is used to make money, for himself and possibly others as well. He is likely an investor in other audio companies as well given his affinity to blend business and research. He's also said that his research striking a nerve in pop culture did a lot to boost his personal notoriety.
Circling back to the 2016 study, it was actually not an experiment he personally set up but a metanalysis of other studies. The "12,000" respondents number looks large and robust but this was not an ad hoc study. It was piecing together and drawing conclusions from the works of others. It was not 12,000 people who specifically sat down and took a A/B test on hi res audio but drawing conclusions from lots of different studies.
Two key conclusions of the study were that untrained respondents were almost universally 50/50 crapshoot guessers while the trained listeners were able to guess correctly up to 60% of the time. They also did better with clips longer than 30 seconds vs short clips where it was much harder to distinguish.
My educated guess is that whatever training people went through to try and pick out the hi res source was likely a non music related tell as I have suggested such as a delay in preloading larger files sizes or a difference in level matching. Either way we are missing a lot of key data about the methodology. We have no repeatability, the robust claim is dubious, the guy who conducted the analysis is in business with brands that would benefit from the success of an obscure format taking off (it largely failed).
All this and you still have a burden to prove that a great DAC can't remove aliasing and jitter to inaudible levels and that there would still be some benefit of a larger bit bucket.
|