Continuing with CDs?

aargh sorry

There are quite a few examples of the redbook layer on hybrid SACDs being quite radically different - different mastering path, more compressed, etc; the players themselves tend to have different output levels for each source, so it becomes difficult to compare. Dark Side of the Moon is frequently cited as an example of a hybrid SACD (maybe the best selling disc of its type?) where the CD layer is worse than the commercially available CD!

There are quite a few boutique labels out there that offer hi-res samples for download, to satisfy listeners' curiosity. It's sort of interesting (sort of) to see what you get when you subtract the CD version from the ultra-hi-res. Last time I did this (a small jazz ensemble, mixing inverted 32/352k with the 16/44.1 version), what was left was a hump of LF (0-70Hz) at least -90dB down, nothing but thermal noise 70-22k (-150dB), and then the resumption of whatever ultrasonic content had been captured by the mics (or introduced as noise by the gear) around -90dB to -110dB, for the rest of the spectrum. Naturally, this sounded like complete silence, even cranked up. But if I played it through VLC (rather than Audacity), I could hear the faint shuffle of the brushed drums. I assume VLC was using the 24/44.1 Core Audio settings on the laptop, and somehow folding the ultrasonics down into the audible region. I can see how this would very undesirable with something like DSD64, cos it's not musical content up there with DSD, it's quantisation noise pushed out of band. Or, y'know, maybe not. People like vinyl ;)

it's not that hi-res magically gives you access to frequencies you wouldn't hear otherwise. it just allows you to use less abrupt filtering techniques. even the most creative filtering techniques for redbook format introduce temporal distortions in the audio spectrum (ringing, or time smear). at issue is not whether you can hear these distortions (some people can, and you can be trained to). the question is whether you can *perceive* them. do they annoy your brain, in some sense? people commonly refer to digital "fatigue," where after a while you get sick of listening. this is such a loaded topic, but there is science to back it up.

bringing up MQA is almost as combustible as bringing up mask mandates, but damn the torpedoes: https://www.soundonsound.com/techniques/mqa-time-domain-accuracy-digital-audio-quality

There are filtering techniques that are inaudible. Yes, fatigue and other non-quantifiable effects are perceived as real. It's the same with people who are allergic to electricity or microwaves or mobile phone signals or wi-fi signals or 5G or whatever the next technology will be. The people perceive it as real, but nobody can prove it actually exists.

Yep, fair enough. There are quite a few different approaches around to stop-band attenuation (you can even select different filters on some CD players) - sharp roll-off, shallow-and-early (HF roll-off), shallow-and-flat-to-20k (but some aliasing). Again, whether one can hear the effects of these different approaches is debatable. But, yes, a higher sampling rate to begin with means that the potentially problematic effects of sharp low-pass filtering are avoided.

xxp

the higher sampling is only needed for recording, though. once the filtering is done, there is zero reason for it to be there for playback. same for bit depth, 16 bit is already excessive for safe listening but it's a sane amount of overkill.

technology wise, they got everything right the first time with CDs but upper management thnakfully were too clueless to anticipate the eventual ease of making backup copies so ever since the late 90's it's been a never ending cycle of forcing not only useless but technically inferior DRM formats down consumer's throats under the guise of "better quality"

braised cod otm

the filtering i'm talking about is done on playback.

i was referring to the LPF to remove frequencies above nyquist prior to ADC, that's the only place where the higher sampling rate makes sense. there's no need for sampling rates higher than 44.1/48 khz during playback, as the aliasing has already been prevented with the LPF before digitisation

the "time smearing" FUD is MQA marketing bullshit, citing MQA propaganda as evidence for redbook's inadequacies isn't going to fly.
at least DSD, hi-rez PCM and other DRM'd nonsense formats, despite the bloat and higher risk of lower fidelity playback, actually can (under ideal circumstances) work as well as redbook audio. MQA is a 100% certified scam

This is a decent article about filter choices in DACs and why sharp filtering is nothing to be worried about:

https://addictedtoaudio.com.au/blogs/how-to/how-to-pick-the-best-filter-setting-for-your-dac

here is an article that shows that MQA does very well in the time domain, with supporting test data. but as i said, MQA, and hi-res sound quality in general, are things that simply cannot be discussed on the internet. the conversations always devolve into "listen to whatever makes you happy." and so shall this one. https://www.stereophile.com/content/mqa-tested-part-1

i suppose i wasn't clear enough: MQA is objectively a scam, no ifs and buts about it. it is pure snake oil

unlike SACD and the like where there's still the "you can't prove other people can't hear things just because you can't" angle braised cod mentioned, where reproducible, rigourous scientific testing of the supposed benefits of hi-res with a large number of listeners is hard to do without bias (individual testing is trivial)* , MQA has already been thouroughly falsified.
it's a lower-fidelity, lossy scheme, as much of a DRM cash grab as any other "alternatives" to redbook except this time there's zero chance of it even being as good, let alone *better* than an already perfect delivery format

* take your hi-rez/DSD file, convert it to 16/44.1 PCM, then convert it back to the original format (exact same specs), then ABX test the original and the double-converted files.
of course, the issue with individual testing is 1) anyone can cheat and post logs to the internet claiming they can hear a difference, and 2) even honest testers might make accidental mistakes and not be aware of it

When I was a kid I had a Walkman with MEGA BASS switch that made low frequencies louder. MQA is like that, but you can't turn it off because it's lossy.

Re: SACD's from pre-DSD digital masters, I was surprised how many SACD's were being done for albums widely known to have digital masters. To be fair, the regular CD's weren't necessarily done from the same masters, especially if they were originally released in the vinyl era - in those cases, rather than accessing the digital master (which may have been too difficult to play back due to obsolescence), they might have played back the vinyl cutting master. (When the Beat remastered their catalog, they had a tough time finding someone who could play back the

But as mentioned, Dire Straits' Brothers in Arms was remastered by MFSL for SACD (all-digital recording) and they even remastered Richard Thompson's Rumour and Sigh, Los Lobos' Kiko and Bob Dylan's Oh Mercy, all of which were recorded in analog but ultimately mixed down into PCM digital masters. IIRC if you can read the DSD files on the Kiko SACD, the data clearly goes off a cliff right when it hits 20 kHz. MFSL hasn't told anyone what they did, but I'm guessing it's not dissimilar to what others do, which is feed the digital master back into the analog domain, then back to SACD. Greg Calbi actually talked about this when he did the 2003 Dylan remasters for SACD - he mastered the original CD for Oh Mercy which was a straight transfer from the PCM master, but it made no sense to him to do that again for the SACD. Feeding it back to analog made sense because he got the idea to use unique analog tools as a way of coloring the sound and getting something that was not only different but also desirable in its own way. You could say this degrades the sound, but arguably ANYTHING done in analog or really mastering "degrades" the sound, it's really more about shaping the sound into something pleasing. So I'm guessing that's what MFSL probably did with their SACD's as well - bring it into the analog realm for mastering and then outputting it again into an SACD master. It doesn't restore info that was never captured by the recording to begin with, but it does add some data that could be preserved better if the master is outputted into high resolution.

FWIW, I came across this interview with Dave Wakeling about the English Beat remasters, which shows the difficulties of remastering early digital. (Their second and third albums were digitally recorded.)

https://www.popmatters.com/162069-special-beat-service-an-interview-with-english-beat-2495824800.html

"One of the nice things about this whole process was it forced us to go back on the master tapes and check on their health. A lot of them were starting to shed if they were on analogue tapes and a lot of had been recorded on the first digital systems to come out, which was a 3M system. Which kind of ended up to be the Betamax, you know it never really made it. So we had to take the tapes to France where there was one 3M machine still operating. So we had to take them there and get them transferred to another format. At least now as part of the process of doing the box set, we have all the masters preserved and if we hadn’t gone through that process in another few years we may have lost some of those songs forever."

I'm just glad that I haven't heard anything lately about Neil Young's feline hearing and his porno player but instead just about how much he hates Joe Rogan. Maybe getting married has exacerbated his selective hearing loss.

i happened to notice this paragraph in the cover article of the new stereophile. it's written by editor jim austin and expresses the thought i was trying to convey above better than i did. it's in reference to a $46,000 transport, but still. Early digital focused largely on the frequency domain. As a result, mistakes were made. The “Red Book” standard for CDs settled on a sampling frequency of 44.1kHz because that was the minimum rate needed to cover the full audible range (you must sample at twice the bandwidth in order to allow “perfect” recovery of the original time series, which gets us up to a sampling frequency of 40kHz) plus a narrow transition band to allow for bandwidth-limiting. But the folks who defined the “Red Book” spec didn’t allow enough room for optimal filters—just sharp, fast ones. Sharp and fast in the frequency domain equal broad and slow in the time domain. At CD resolution, you can get near-perfect frequency response or good time-domain performance, but you can’t have both.
https://www.ch-precision.com/wp-content/uploads/2022/02/Stereophile-March-Issue-2022-CHP.pdf

Too bad they still can't prove it in a listening test.

depends who "they" are and what the test is. trained/professional listeners do much better in blind tests, and blind tests aren't a great way to pick up these kinds of effects, which for most listeners register as fatigue over long intervals of listening.

"They" are everyone who claim to hear a difference without doing a proper double-blind test. Trained and professional listeners do better for the obvious reasons, they know what to listen for. Double-blind tests are the most reliable method to hear audible differences and that's why they're rigorously used in every field of audio industry.

Most of the sounds people hear are compressed audio, whether it's TV, radio, phones or streaming media, the codecs are developed and tested with double-blind tests to make sure they're as transparent as possible. And yes, you can't test fatigue or other non-measurable effects, that's how healing crystals work.

i dunno, i am glad i don't have to do a double-blind test every time i claim to hear a difference. but basically that sounds tautological. you develop a test that you know untrained listeners will fail, then they fail it, and the conclusion is: aha! proof positive that cds are perfect!

whatever, the grateful dead guys use 24/96 as their standard, and they're always at the forefront of everything.

if there's anyone to trust about a fatiguing listen

haha well since their concerts are 4 hours long, they would be the ones to know.

> you develop a test that you know untrained listeners will fail

no trained listeners ever passed a properly designed ABX test comparing redbook against hi-rez. it's not only untrained listeners that fail, everyone fails

you claim to hear a difference because you're expecting one, that's why blind tests exist

I've never encountered a self-professed audiophile who seemed to care much about or have heard very much music. Certainly not who had wide and deep listening interests.

Results showed a small but statistically significant ability of test subjects to discriminate high resolution content, and this effect increased dramatically when test subjects received extensive training. This result was verified by a sensitivity analysis exploring different choices for the chosen studies and different analysis approaches.
https://www.aes.org/e-lib/browse.cfm?elib=18296

i really don't get why people are dead-set against exploring other digital options that provide a better listening experience. why does science have to stop in 1985?

might as well ask why do people still believe in gravity or why the earth is flat, why stop indeed? why not post sales brochures as evidence of scientific progress?

re: the AES meta-analysis, what a shocker that a review would include results from improperly conducted studies, then discover there's a difference after all

https://hydrogenaud.io/index.php?topic=112204.0

hey now that's just not playing fair. if you mean by "sales brochures" the stereophile review of the ch precision transport, i happened to grab it from the vendor site because they were proud of the review and posted it in full, and it is not available as free content on stereophile's site. but if you want to read it in stereophile you can go buy it off the newsstand. and yeah what a shocker that the aes, the largest professional organization of audio experts, would post an analysis and people would then snark about it on some internet forum?

by sales brochure i meant stereophile itself, not where the pdf was hosted

so SP is something to be trusted and HA is just "some internet forum", got it

xp It's not a scientific paper or peer-reviewed therefore sadly the scrutiny is mainly only found on internet forums.

i'd like to see archived usenet discussions on qsound.

i think that paper has been peer-reviewed. it appears in the journal itself (JAES Volume 64 Issue 6 pp. 364-379; June 2016). fully expecting a hydrogenaud link to someone who says peer reviewing doesn't mean anything anymore. there's no way out. i had not gleaned that you meant stereophile magazine itself is a sales brochure. the layers of snark run so deep here i'm out of my element. i'm gonna go back to the blue oyster cult thread where they like me.

hey tsf, no snark intended on my part. i'm just a wisearse that doesn't have anything of substance to add to this discussion. but i like reading it either way, so thanks for posting.

(also qsound marketing was funny.)

i really don't get why people are dead-set against exploring other digital options that provide a better listening experience. why does science have to stop in 1985?

― Thus Sang Freud, Saturday, February 12, 2022 9:45 AM (two hours ago) bookmarkflaglink

I don't have the faculties to parse the science, but I do sometimes think it's odd when people have such a resistance to the idea that music formats and playback technology can't improve over time just looking around at.... literally every other technology in the world

haha xp austin my comment was not directed at you and your comment was funny.

Isn't hearing, for most of us, getting worse the older we get?

Won't most of us, by the time we can afford better gear, be less able to appreciate whatever difference it makes?

some of this is cognitive, is the point i've been trying to convey. it's not just the sensor itself but what the brain makes of the info. if we were merely talking frequency response, then yeah as we get older our sensors degrade. but the point of hi-res audio is not to increase the frequency range. that's one thing that cds are good at. but they're not so good at timing. humans are very sensitive to timing, from prehistoric days. if hi-res audio can fix the timing issue, then digital audio becomes more transparent, and deep listening becomes that much easier, even if your hearing has degraded.

Isn't hearing, for most of us, getting worse the older we get?

WHAT??

ha.
thread delivers the oldest, but still brilliant, joke.
i have tried to keep up with the audiophile detour but the simple fact is : cds are great.
yeah, i know i downgrade the quality when i rip to 320 for my sonos archive (originally ripped at 256, redoing the collection killed the optical drive in my laptop, but was worth it), but, i have decided to be happy with the compromise and just kick back and enjoy having instant access to all of my collection.
of course, should my NAS drive die, then i have it backed up, but should the worst of the worst happen, then i still have the original cds (for 93.6% of the collection).
albeit scattered semi randomly all over the house/attic making it very hard to find a specific cd as and when the urge should kick in.

tsf could you please define what “timing” means and how CDs lack it?

This is kind of what Bob Stuart of Meridian has been claiming to "solve" with MQA - a sort of time-domain "deblurring", but without ever really being specific about how it's achieved or what "deblurring" really means or how one could possibly eliminate all the phase-shifts / timing issues that accumulate in a multitrack recording chain. The articles I've skimmed on this suggest that humans are sensitive to time-domain inaccuracies of the order of a few microseconds; but redbook digital has time-domain resolution in the order of picoseconds.

Audio tech has improved, and continues to improve, in all kinds of ways (I think £250 speakers / £40 earbuds these days are pretty amazing compared to what that bought you 30 years ago; power-efficient class D amps are way better than they were, etc), and even CD-level digital is better, as ADC/DACs have improved. But hi-res digital isn't adding much to this progress as far as I can tell.

Isn't hearing, for most of us, getting worse the older we get?

How can I miss hearing quality? It's so obvious but it completely passed me by.

How many of you regularly go to concerts and wear ear protection every single time? I'm guessing not many, or if you do only sporadically, because I rarely see anyone else wearing ear plugs or ear monitors for protection, and that will take a toll on hearing.

I have really excellent hearing that's held up over the years, and I have to thank a classmate who sadly was going deaf in one ear. I remember because the first time I ever saw ear plugs was when I walked into our very first school band concert and noticed a small pile of ear plugs in cardboard packets sitting on the percussion table. A classmate also in that section told us he brought them for us, explaining he already had bad hearing loss and that we needed to be careful. We got along very well and I had no reason to doubt him, so I took a pack and wore them for the show. Eventually I got a jar of them at a drugstore and made a habit of wearing them. I was especially conscious of doing so because it's hard not to notice that practicing drums gets really loud. Many years later, not long before I moved to NY, I was with a friend who was fiddling with a device that played 20k tones. It was sort of a hearing test in a way because he heard nothing and neither could an acquaintance of ours who went to concerts basically every week with no hearing protection. In fact, the same guy had trouble hearing lower frequencies (I remember it being 16k but I'm reluctant to go with that because that's pretty bad hearing loss). On the other hand, myself and a friend's younger sister who was there could hear the 20k tone perfectly. (Supposedly women and obviously younger people typically display better hearing at higher frequencies, which is why I bring up her gender and age.) I have no doubt that wearing protection has made an enormous difference. Also it's already been widely reported that doctors are finding younger and younger people with the type of hearing loss they expect from the elderly, and it was often chalked up to portable devices being played way too loudly. I don't doubt this either - it's a big reason why I got ear monitors that sealed out outside noise. (I don't have to crank up the volume to hear a bass line over the rumble of the subway.)

So with all that in mind, if you're already at that point where the upper frequencies of your hearing has dulled, the advantages of high fidelity have already been greatly diminished. You even see this in bad mastering - it's a joke among some mastering engineers that when they need the approval of an older listener (say someone in the band), it's inevitable they'll have to boost the upper frequencies because that spot in their hearing has inevitably been decimated from decades of performing without any hearing protection. There are countless remasters out there that have a shrill, piercing EQ curve, and it wouldn't surprise me if that was the usual reason for it.

OK could you define what “time-domain” is, and perhaps as a bonus consider how time-related phenomena in the microsecond range would compare to head position and movement? Sound travels 0.3mm in a microsecond. To be sensitive to “errors” (compared to what?) on that scale you’d need headphones screwed directly to the skull, otherwise breathing, head position, etc. would absolutely swamp any “time-domain” “smearing” effects.
Sorry, just extremely weary of marketing hype disguised as pseudoscience.

sorry that was a reply to the previous

Michael Jones - see also the Gas thread for "old techno guy can't hear the audio errors in his music" roffles

wait sorry that was to bird

Yeah, I couldn't believe that Gas album! One occasion when streaming beats physical media, I suppose; it was fixed and replaced on all the platforms, but the folks who pre-ordered the CD weren't so lucky.

usually frequency domain is the concept people have trouble with. time domain is straightforward. you know that effect they put on robert plant's voice in the spacy middle part of "whole lotta love," where you sort of hear him said "way down inside" off in the background *before* he actually says it? that's a macro version of what's happening at a microscopic level when you have ringing in the signal. the notes aren't happening precisely when they're supposed to. there are little artifacts preceding (and also coming after) the sound.

Not sure that's really the big deal it's been made out to be. With DSP, recording plug-ins, or with psychoacoustic lossy codecs, certainly it's a real problem - but with properly bandwidth-limited 16/44.1 digital playback, it just isn't there. Becomes more of an issue with heavily compressed / clipping modern pop, where "illegal" waveforms are being fed through the reconstruction filter.

Good article here:
http://archimago.blogspot.com/2018/01/audiophile-myth-260-detestable-digital.html