Dr. AIX's "Debunking" Hi-Res Audio blog post

[humprof]

Dr. AIX's latest blog post/newsletter delivers sort of an abstract/executive summary of a paper he'll be delivering at the AES convention in October (however that's gonna take place). In the paper, he reports the results of a study he undertook on whether listeners can actually perceive a difference between RBCD and hi-res audio. (Short answer: no. But the long answer is more interesting.) Since people have debated this question on QQ before, sometimes hotly--and since my own handle on some other forums is "ponoskeptic"--I thought I'd bring it to people's attention...
https://www.realhd-audio.com/?p=7037
By the way: the final sentence of his summary may answer the question that @Sal1950 asked in another thread:

hi-res doesn’t matter as a final delivery format. However, it does matter while Bill Schnee, Cookie Marenco. or I are making new recordings. We can all be right. And audiophiles should understand the facts.

 

[J. PUPSTER]

Many factors goes into the sound that the end listener will have to experience, as well as their own gear, room acoustics (headphones), listening ability (and prejudices); but Mark @Dr. AIX is welcome to stop in and chat anytime AFAIC. I'm still not sold on what I hear from one format and resolution to the next, but it sure seems like if I play an RBCD and follow it up with something that's 192 khz / 24-bit , I believe I can hear it (he was comparing 96 khz / 24-bit though). But then again, I've never done any real blind testing of this

 

[EricKalet]

Mark Waldrep is a very smart man and knows mastering and high res. His argument regarding high res, and rightfully so, two main themes: provenance of the music and the quality of mastering to digital. A few inferences can be made: if music is recorded at 96/24 then it makes sense to master to 96/24 for physical or streaming playback and you get that 96/24 quality. If music is recorded 44.1/16 than it makes no sense to master up to 96/24 because scientifically there would be no audible benefit.

However if music is recorded at 96/24 and mastered down to 44.1/16 in most cases, human hearing should have great difficulty distinguishing an audible differnce in quality during playback of the same music at the 96/24 recording or the 44.1/16 recording.

 

[jimfisheye]

I hope it's at least intellectually honest. So many A/B/X tests using a stepped on loud and treble-y master that would sound the same on a wire recording machine when this stuff comes up.

I think there are a couple fair points to be made though. (Based on my experience FWIW.)
Cheaper DACs used to be significantly poor sounding at SD sample rates. Not so much of a gap anymore. There's also the irony that the higher end AD and DA converters sound very nearly or truly just as good at SD sample rates as they do at HD sample rates.

Early generation sample rate conversion was pretty damaging. Especially converting SD to SD (44.1k < > 48k). 20 years ago, going out DA -> analog connection -> AD to "convert" sample rates was always cleaner. That's changed with algorithms like SOX.

The volume war CDs with the bionic treble boosts...
We're SO far away from anything to do with format limitations with this. You can't really even evaluate that variable. That's not the sound of 44.1k or even 16 bit!

I think edge cases can still be a thing and generation loss can stack up. (Generation loss being analog domain, digital resolution based, or digital conversion or transcoding based.)

24/96 is bulletproof I think. No matter what previous generation loss a previously damaged recording might have, 24/96 with proper converters and analog stages captures the full recording. No more edge cases. No matter what.
But I think 24/48 would be a fine format and not cause any damage 99.999% of the time. The limiting factor will still be the quality of the mix and the handling in mastering. Lots of crude mixes out there. Especially the pop stuff right now.

I've heard all kinds of generation loss artifacts over the years. I've been a card carrying member of the audiophile police. I upgraded sample rate clocks back in the day (remember those Aardvark units?) and heard a big difference. All the conversion stuff I've mentioned. Bought into Apogee back when they were just a stunning upgrade. Copying an already troubled analog recording and finally reaching a stage to be able to make a copy that sounded exactly the same with no further snowballing of quality. (I keep mentioning the troubled generation loss recording because it's a ringer of an edge case and calls out weakness. Of course a pristine recording will show no damage short of gross operator error copying from point A to B.)

And I really think we've evolved to where any decently pro converter units will deliver at 24/48 and you would not pick out a difference in a million dollar listening room. I'm sure there are still some muddy AV receivers out there with crap converter circuits that perform better with HD program. But you know what? You can upsample those files to 96k with SOX and then hear them more true to the master. (You'll actually find some receivers that upsample as their standard feature to achieve better sound.)

My 2c.