[bigshot]

I've found that with the couple hundred SACDs and other "hires" formats I've bought, the odds of getting something that has better sound quality than the standard CD release is about 1 in 4. Half the time it sounds pretty much the same. 1 in 4 sound worse. Pretty much random. Multichannel has the potential for improving sound beyond just the quality of the mastering, but the quality of multichannel mixes vary too. I haven't found any correlation between hires and quality of sound at all.

 

[KeithEmo]

I'm inclined to agree with you..... some remasters sound better, some worse, and there seems to be little correlation with whether they happen to be high-res or not.
(Although, personally, while I like surround sound for movies, I tend to prefer stereo to surround sound for music.)
And, yes, I agree that the mixing and mastering make a FAR greater difference than the sample rate.

I've found that with the couple hundred SACDs and other "hires" formats I've bought, the odds of getting something that has better sound quality than the standard CD release is about 1 in 4. Half the time it sounds pretty much the same. 1 in 4 sound worse. Pretty much random. Multichannel has the potential for improving sound beyond just the quality of the mastering, but the quality of multichannel mixes vary too. I haven't found any correlation between hires and quality of sound at all.

 

[KeithEmo]

You're probably right. And, after hearing a few really well mastered CDs, it's very disappointing how bad the majority of the discs produced lately actually sound.
(Of course, they may sound exactly how the engineer wanted them to sound, or how he or she imagines the audience wants them to sound.)

It's also pretty obvious that re-masters have always been, and will continue to be, a viable way to sell more copies of albums that people already have. People have always been willing to buy "new and improved" copies of their favorite albums, and, if high-res, or MQA, provide one more selling point they can use to sell a few more, of course they're going to do so. The actual COST of storing a 24/192k copy instead of a 16/44k one is trivial for the vendor, as is the cost of the extra bandwidth it takes to download 1 gB of extra data these days, so, if it sells even a few more copies, then it's going to be worthwhile for them.

(I would remind everyone, however, that remasters are nothing new... Dark Side of the Moon has been remastered somewhere around two dozen times.... and I assume people still buy the new ones. Studios have a huge investment in their music library - so they aren't going to stop trying to find new ways to make money from that investment.)

Another interesting thought is that, if a vendor sells copies of the same album at 16/44k and at 24/192k, it might actually be in their best interest to make sure that the 16/44k version sounds audibly worse - even if that entails using a different mix or using an inferior piece of software to create the "cheaper" version from the "premium" version. (I haven't heard anyone accused of doing so, but it might be interesting to see if, in specific cases, the 16/44k version someone sells is slightly different than what you get when you down-sample the 24/192k version they sell.)

I would also advise caution when assuming that all programs do an equally good job at converting between sample rates.
There is a website that has published comparisons of how different programs perform when doing sample rate conversions.
It's interesting to note that, while many do an exemplary job, many do a rather poor job.
(Interestingly, for example, the CS6 version of Adobe Audition does an excellent job, while the CS6 version of Adobe Media Encoder does not.)
It's also informative to compare both sweep spectra and impulse response (note that most apodizing filters have a lot of aliasing).

I'm not interested in getting into a discussion about which flaws are audible.
(Or why, when the technology is obviously widely available, anyone would still do it badly.)
However, it makes sense to me to choose a program that technically does a good job, and avoid software that does a technically inferior job.
(And it makes you wonder which CDs were, perhaps, converted using one of the converters that performs very poorly.)

That website is: http://src.infinitewave.ca/

I'm not here to tell you how to spend your money. That's your own business. If great remasters are sold only at 24/192 kHz and you want them of course you buy them, but if they had them available at 16/44.1 kHz too, you probably would find them equally great. Maybe the idea was that it pays off to remaster if you can sell them at higher cost because it's 24/192 kHz?

 

[danadam]

Another interesting thought is that, if a vendor sells copies of the same album at 16/44k and at 24/192k, it might actually be in their best interest to make sure that the 16/44k version sounds audibly worse - even if that entails using a different mix or using an inferior piece of software to create the "cheaper" version from the "premium" version. (I haven't heard anyone accused of doing so, ...

Well... there's this:

Not so much actually record but master differently, certainly. In fact, way back earlier in this very thread I remember an exchange I had with a representative of a distributor, I can't remember which distributor but I think it might have been HD Tracks or Linn, where they stated they do (or did at that point in time) routinely make their 16bit versions poorer quality! Of course, that's not exactly what they stated, they stated that they make their recordings as good as possible and that in the case of their 16bit versions they add a significant amount of compression. This is because, they say, many of their clients then lossy encode their 16bit version for use in portable devices where more audio compression would sound better. Now that is a reasonable/acceptable response but when I basically said fair enough but instead of distributing a more highly compressed 16bit version and a far more expensive "HiDef" version, why not just sell a second 16bit version but without the additional compression (which would be audibly the same as the HD version). No response, the silence was deafening!

 

[bigshot]

I have a Rolling Stones SACD where the redbook version isn't even the same mix as the SACD layer.

 

[old tech]

I have a Rolling Stones SACD where the redbook version isn't even the same mix as the SACD layer.

Interesting that it is a different mix, however it is not uncommon for the CD layer of a SACD to be more compressed. The rationale of the labels is that the CD layer is better for everyday listening (eg in noisier environments) while the more dynamically mastered SACD layer is for quiet listening rooms.

 

[pinnahertz]

1.
I'm sorry, but, yes. If the speaker has output past 23 kHz, and I can measure that output with a microphone at the listening position, then the tweeter is indeed "delivering ultrasonic energy to the listener" (assuming we're referring to 23 kHz as "ultrasonic").
In the practical world, if we can measure something, and the measurements show it's there, then it's there.
(I would agree that, if it's 20 dB down at 25 kHz, it probably won't be doing anything useful at that frequency.)

2.
Here you go.
Here's a link to our Emotiva Stealth 8 studio monitors.... (they're about $900 each - so cheap by audiophile standards): https://emotiva.com/product/stealth-8/
They're rated at:
30 Hz to 23 kHz + / -1.75 dB
28 Hz to 32 kHz +0 / -6 dB
And here are the graphs: http://cdn.shopify.com/s/files/1/0201/8878/t/2/assets/Stealth8_Plot_Graph.pdf
And, yes, their response drops off rather rapidly off axis at very high frequencies (but I usually sit in front of mine).
And, yes, those are real measurements.

3.
As for microphones..... I agree entirely...... and it's up to a whole bunch of people what microphones get used, and where the mixing engineer sets his (or her) filters and EQs.
My only goal is to make sure that MY equipment isn't the limiting factor - so I need to be able to handle anything that they MIGHT include.
(And, yes, in engineering, it's pretty standard to include SIGNIFICANT safety margins to eliminate the possibility of this happening.)

And, yes, if the mixing engineer decides to use a microphone that only hears up to 15 kHz, or even to apply a 10k low cut filter, that's his business - because he's PRODUCING music.
However, if it isn't there because MY equipment failed to reproduce it, then my equipment is broken - because it has failed to REPRODUCE what the engineer put there.

4a.
As far as I know the "content" on vinyl CAN extend to about 60 kHz (I'm defining "content" as "retreivable information").

4b. I read that the pitch stability on some of the Grateful Dead albums reissued by HDTRacks was corrected by locking onto the residual 80 kHz carrier tone from the erase bias - which is still recoverable on tapes recorded on certain master tape machines.

4c. I agree that I have no plans to either use a vintage click remover or restore any antique tapes.....
However, having extra information that I don't ever use is at worst useless, while not having information that I turn out to need later could be tragic.

5a.
My point is simply that, if you recorded a CD-4 record directly off the stylus, and later decide to feed that signal to a decoder, it won't work if those frequencies weren't recorded.
This would obviously be a poor method for archiving four-channel content on CD-4 records.

5b. And, again, if you want to claim that you have AN ACCURATE RECORDING of that Cd4 album, then it should be there.
(Or you must concede that you have saved a "processed" copy of the content rather than an accurate copy of the original.)
You have simply decided to set YOUR goal to exclude the requirement of saving all of the out-of-band content in its original form.

6.
I'm not forgetting that at all.
BY DEFINITION, whatever comes off that mastering console is "right".
BY DEFINITION, the job of a "high-fidelity reproduction system" is to reproduce whatever I play through it accurately and completely.
I absolutely agree that we have no provenance, and that it's a problem.
However, while I can't fix that problem, I can at least make very sure that I don't cause unnecessary damage or alteration.
I can only provide quality assurance on the parts which I control.

7a.
If you're really trying to say that "damage is OK" then I guess we'll just have to agree to disagree.
In most situations, if a perfect example is available, it will be valued more highly than a damaged one.
That doesn't rule out the possibility that there may be no undamaged examples available, or that a certain damaged example might not actually be considered better than a perfect one for other reasons.

7b. However, I don't think you'll find many dealers of legitimate antiques who would suggest deliberately breaking a vase, and then gluing it back together, as a way to increase its value.

7c. Of course the price will be DIFFERENT if it's damaged.... and, of course, if you prefer to pay less for the damaged one, then that's your choice.
(Try putting up two ads on eBay - of the same vase - describe it in one as "perfect" and in the other as "broken - but with a perfectly invisible repair" and see which one goes for a higher closing bid.)

7d. It occurs to me that this provides a perfect analogy to how *I* feel about both lossy compression and storing music in any format that is designed to carefully avoid storing any level of quality that isn't strictly and provably necessary.

7e. To ME, that would be as if I had an undamaged vase which I needed ship, and someone were to suggest: "If you break it into pieces it will fit in a smaller box and be cheaper to ship. We can glue it back together at the other end and
none of our customers will be able to tell the difference.... so what's the harm?"

8a.
The reason I bring up lossy compression is the same reason that your favorite studio does NOT store their masters using it.
In order to "work", and deliver a "virtually indistinguishable" copy of the original, virtually all lossy compression relies on a whole bunch of assumptions and given conditions.

8b. Feel free to suggest that, "in order to hear the noise floor on a CD I would have to turn my system up so loud that the loud parts would deafen me". However, you seem to be ignoring the fact that I might decide to turn the volume way up on a quiet part to hear what one of the musicians mumbled under her breath.....
Are you suggesting that it is INVALID for me to do so?
Or are you simply suggesting that YOU wouldn't do it?

8c. JPeG is a remarkably effective lossy compression for images; but NOBODY sane would use it as an archival format.

8d. The reason is that, while JPG images may be "visually perfect copies of the original" under certain circumstances, that isn't true under ALL conditions.
Try boosting the contrast, or looking a little too closely at one tiny spot on the picture, and you often notice the flaws.
And, if you were to try and edit that picture, it's a virtual certainty that you'll "run into the limitations" caused by discarding all that data.

8e. Of course, in real life, we always have to choose a set of requirements that suit US.
Someone who favors 24/96k as a recording format will be quick to point out that reducing the sample rate to 16/44k is obviously a form of lossy compression.... You are discarding information which cannot be gotten back. I guess it also qualifies as "a perceptual lossy format" since you decided what was OK to throw away based on what you can hear - and what you expect me to be able to hear.

9.
As for ANALOG masters, since there's a format conversion involved, there is no direct comparison.
You may prefer that the tape hiss be omitted; someone else may insist that it be reproduced ACCURATELY.
(Just as, in video, some people prefer a smooth background, while others delight in figuring out what type of film was used by the original camera crew by examining the shape of the noise grains.)

Back in the analog days, it was IMPOSSIBLE to own a perfect copy of an album, because every copy process added a small amount of error.
Digital technology now makes it POSSIBLE for every customer to own an actual IDENTICAL copy of that original digital master.
To me, it seems like some sort of blasphemy to throw that opportunity away just to save a few bytes - or a few dollars.
It it seems as if I'm not overly concerned with whether they difference is "audible" or "significant"... then you're right.
I simply see no reason to forego "perfect" and go out of the way to look for an alternative which is "not perfect - but I'l never hear the difference".

I also do find it very sad that all of the attempts to claim to provide provenance have turned out to be either overreaching or just plain impractical.
I think it would be great if I COULD really buy a file knowing that it was a legitimate bit-for-bit copy of what came off the mastering engineer's console.
(Clearly the fact that so many companies promise it - even though they routinely fail to deliver on their promise - suggests that a lot of people agree.)

I'm also going to disagree with you on your definition of the word "practical".... simply because it is a word associated solely with personal opinion.
You seem to consider 24/96k to be "not practical".
I define 24/192k as quite practical - because almost all of the DACs I currently own support it (but I consider 32/768k to be a bit impractical with current equipment).
(Note that nothing was said about "necessary" - which is a different value judgment.)

1. That's exactly the kind of warped logic I'm talking about. It's like defining "any"="all".

2. Nice. I'll admit to being impressed, except for the even slightly off-axis response, which means this works only for one seat. And except you said "there are plenty of speakers", and you've cherry-picked an example. You also said they measure flat past 25kHz. I don't consider -10dB@30kHz "flat past 25kHz". No matter, you've found an exception, not the norm or typical.

3. NO relation to reality here. There's no "production" going on above 20kHz. How could there be? Do you think all control room monitors are flat to 30kHz, or that engineers hear above 20kHz? The 10kHz high cut filter example is also ridiculous, completely over-reaching reality. The actual question should be: how high is actual music content? That's content that is actual music, not distortion products caused by transducers.

4a. What's the 3rd harmonic of 20kHz? How about the 3rd harmonic of 8kHz? Are you aware that harmonic distortion in the vinyl system is HUGE? See, this is the real problem, the assumption that 60kHz coming off vinyl or even 25kHz is somehow part of the original acoustic event. Not very likely, especially considering they were mostly mastered from tape, which rolls off quite quickly above 20kHz. The nonlinearities in the total system are teutonic at high frequencies. The distortion characteristics jump up quite quickly as a function of level and frequency because of the physical limitations in the geometry of the groove, cutting stylus and reproducing stylus. For example, you can't have a cutting stylus velocity that exceeds a physical maximum without the stylus's rear facet scraping the groove wall just cut. A similar physical limit exists in playback. Given that, it's far, far more likely that what's going on is the result of distortion mechanisms...many of them. What's wrong is, since you don't have the original acoustic event PLUS an ultrasonic-capable microphone system to evaluate it with, you make the assumption that it was original. But that's just not true.

4b. Another completely unrelated point. The bias signal used in analog tape is several orders of magnitude hotter than the actual audio signal. It literally operates for the purpose of linearizing the recording capability of an otherwise extremely nonlinear medium. Tape must have very high coercivity and retentivity to store information, which results in terrible linearity and very high distortion. The bias signal enables acceptable distortion by carrying the audio signal way past the nonlinear portion of the magnetic characteristic of the tape, while still using a very magnetically stable medium. A bit of residual bias signal recorded on tape has been used by the Plangent process to effectively correct time-base errors, but that process involves special tape heads and extensive DSP to recover the bias signal and be able to derive a time-base signal. None of that has anything to do with the actual HF response of magnetic tape machines! In fact, that same bias signal tends to erase HF audio as it is being recorded, and the level of bias has to be optimized for each tape type as a trade-off between minimum distortion and equalizable HF response. The optimum usually results in flat response to 20kHz, then a rather rapid roll-off above that as bias erasure rate exceeds the capabilities of record equalizers to compensate for it. Once again, an inappropriate example.

4c. In some circles this is recognized as a manifestation of OCD. In others, a full understanding of what those signals actually are will help make a reasonable capture decision.

5a. Yes, that would be a very poor method, and completely the wrong way to do it, as it would take post processing to recover the discrete 4 channels that doesn't exist.
5b. You've taken the discussion over to archiving a CD4 record, and that has nothing to do with the HF response or content of an LP. A CD4 record was a special case, full of flaws and limitations, and it was a commercial failure. The performance of the CD4 system was compromised at best. I have no idea why anyone would want to archive the content of one of those records. This is of course apart from the fragile nature of the HF carrier, which literally could be worn beyond usability with normal play. If the recording is at all important it likely exists in some other more usable form, like the original analog tape masters. Any attempt to even play a CD4 record would be so far inferior to the masters as to be pointless, especially from an archive point of view. The comparison of CD4 with the capability of vinyl to record and reproduce ultrasonic content is meaningless even if you cherry-pick a specific example.

6. Yes, and that quality assurance is admirable. But what would be better is a full and complete understanding of the content, which isn't what's going on here. It seems that you're saying "We don't understand what's really there so over-capture so we don't miss anything". Well, we do understand what's there. It's not even that hard. And that understanding changes what we need to do.

7a. The problem is that a perfect example is often suspected of being a fake or reproduction. A classic case is old metal and porcelain signs. They are so hot right now that they've gone full steam into reproduction. And the reproductions are very, very good! However, they are also easily recognized as being far too perfect, so current reproductions include actual simulated damage. Thats why the actual real old version, with all it's damage, is worth 10X that of a repro in much better condition.
7b. That's just silly, and I'm just slightly insulted at the insinuation.
7c. No, back up one step to "try putting up two ads on eBay of the SAME VASE". First challenge is right there. Then of course if you do have that, and one is damaged and visibly repaired it will be worth much less. That's not my point at all. Have you failed to see my point, or do I need to go through this again? Your example is not appropriate to this discussion at all.
7d. And hold it again. Now you've equated all lossy compression with being by nature of lower quality than the original. Sorry to use your own warped logic, but that cannot be a true statement at all. If there is lossy compression that is indistinguishable from the original, then we must conclude that even a lossy codec exists that is capable of handling all material without audible loss. It's just a question of how much bit rate reduction is applied and how. No, the analogy still doesn't work.
7e. Completely wrong, yet again. A vase cannot be broken into pieces and reassembled without any visual damage or repairs because the damage is random and not specifically chosen to be the type of damage that is invisible once reassembled because the cracks are masked somehow. If you understand how lossy compression works, (and it's not a binary thing at all!), it's data reduction is based on psychoacoustic masking properties which are both variable by parameter setting, and variable dynamically based on content. There is no parallel to breaking a vase.

8a.Yes, of course. Are you implying the assumptions are incorrect?
8b. You've moved over into creative writing now. I don't believe I've mentioned your playback volume at any time, or anyone else's either. I never said anything about playback conditions being valid or invalid. Please don't make up things and attribute them to me.
8c. You've just labeled thousands of photo editors "insane". JPEG image compression is, in fact, and has been a format that is extensively archived. Not necessarily by desire or design, but if an image was originally shot that way, and in the early days of digital photography nearly every image was, then it could remain archived as the original file with no loss. Yes, there are millions of archival jpg images in storage, just as there are millions of lossy film negatives in storage. Here we go again into another inapplicable analogy....
8d. JPG compression is completely variable, and the degree of loss is a choice. What's far more important in image quality is bit depth and total resolution. I have a professional DSLR that shoots JPG and RAW images, both at the same resolution, and the resulting file size (the degree of jpg compression) can be chosen to be nearly identical to the raw file. If the exposure is correct, there's no visible difference. RAW gets you more control of the extremes of exposure because it's essentially a copy of the sensor output, but that's not because of the lack of lossy compression, a RAW file is a completely different beast. A TIFF version of the RAW image would be as limited in contrast editing as a jpg, yet the TIFF is also uncompressed and is not considered lossy. However, Camera RAW is not a "release" file format, as each camera has a different profile and format that are not universally compatible. There is no direct parallel in audio. You need to leave off from analogies, they aren't working.
8e.Incorrect! Lossy compression reduces data based on psychoacoustic masking. Resampling doesn't. If the total audio bandwidth and dynamic range of a given recording can be represented by 16/44, then resampling at 24/96 recording to 16/44 is lossless! No data that represents the original recorded signal is lost because no data actually uses the total dynamic or frequency range in the first place. Resampling can be destructive if the resulting resampled version must exclude data that describes the original, but that's not always the case in audio. Resampling is not perceptual. You are discarding information that may not be gotten back, but if that information only describes noise and not the original signal, it is not a loss to discard.

9. I have no issue with capturing the original in as complete a representation as possible, and distributing that version. I have big problems with the general assumption that high sampling rates and bit depths result in categorically better sound, and question strongly the value of ultrasonic content over 20kHz and the ability of the average or even high-end user to get that energy to his eardrums (much less to actually hear it). I do think there are cases where ultrasonic energy, real or distortion products, can cause problems in devices not capable of handling them without distortion. I do believe there are cases where the process of band-limiting audio by the use of certain devices results in intermodulation distortion that can be folded down into the audible region making wideband audio sound better for a reason other than the ultrasonic content. I also believe that those cases are fewer today than say 25 or 30 years ago, and the real solution is to test for high frequency intermodulation and deal with the cause rather than to band-aid a solution either by passing a wider bandwidth or limiting ultrasonics by filtering. Only a few decades back there were audio products that included ultrasonic filters to prevent those signals from wreaking havoc in other devices. Hopefully today those devices are few, but I doubt they're completely gone.

 

[KeithEmo]

The problem is that you're making a straw-man argument there.

Yes, IF I WERE TO SET MY VOLUME CONTROL AND NEVER MOVE IT, then, in order to hear the noise floor on a CD, I would have to be playing it so loud that the loud parts would damage either my hearing or my playback equipment.
However, like many people, I do occasionally turn up the volume on the quiet parts to hear something better.
And, sometimes, I even get up from my listening seat, and walk over near the speaker..... at which point that nice low noise floor suddenly becomes audible.
Now, I wouldn't claim that we all NEED a 200 dB S/N ratio, just in case we crank it up 100 dB higher in the quiet spots.
However, that's one of those reasons why we include things like safety margins whenever practical; just in case we decide to.

You suggest that "we don't need" the stuff we won't normally hear...
I say that, while we may not NEED it, I PREFER not to hear background noise when I walk by the speaker, or to hear hiss with a less annoying spectrum.
Likewise, even though the highest speed limit I've seen is 70 mph, I prefer to have a car that can do 100 mph - just in case I need to pass someone who's speeding.
(It's also why people pay extra for sports cars that can go 150 mph, or can accelerate to 90 mph faster than mine - because, TO THEM, they're willing to pay extra for that extra safety margin.

I think it's always going to come down to what we consider "normal usage" - which is were we disagree.
I's also true that, in most situations (when you're looking at the picture at a "normal viewing distance" in a "normally lit room"), a JPG picture will look just as good as an uncompressed image.
However, I download a lot of pictures, and I run into situations quite often where I want to zoom in on a section of one, or turn up the contrast......
And, when I do, if all I have is a JPG, I often end up running into the limitations of the reduced quality, where my results would have been far better if I'd had a better quality non-compressed copy.
(I decide the picture is a little dark, so I turn up the brightness, and, suddenly, those "invisible" compression artifacts are very visible..... )

Again, I may agree with you that I USUALLY don't need it, and that it USUALLY won't make a difference....
But I'd rather pay a tiny bit extra for something I don't need 95% of the time than end up wishing I had it the other 5%.
And, yes, when I take pictures I ALWAYS use the lossless setting......
Because, like with audio, storage space in cameras is cheap these days, and I never know when I might need the extra quality or safety margin later.
I'd much rather throw away 50 tB of extra space on the pictures that turn out not to matter than find out that I screwed up one that DOES matter.
(I UNDERSTAND all the people who chose not to buy a camera that can take RAW pictures, or who prefer to save the space.... I just don't agree with their priorities.)

On that note.......
I've purchased quite a few albums from CDBaby (they offer "small volume productions" online).
Up until now, they have offered their content on physical CDs and/or downloads... and their downloads have always been available as lossless FLAC files or a variety of other formats.
(They charge a few bucks more - plus shipping - for the physical CD versions.)
They have recently announced that, while they will continue to offer actual CDs, they will soon only be offering downloads in lossy formats.
(They cite, as an excuse, that they can "ensure the provenance of a CD but not a file". Since they sell CD-R's, created from files anyway, I consider this to be a specious excuse.)
Obviously I won't be purchasing the download versions any more (just as I avoid purchasing music from iTunes)...
I guess we'll see how many other people agree (if we ever hear).

Why not just do a little googling and find out what additional benefits higher bitrates and sampling rates offer, then compare that to the thresholds of perception and you'll have your answer.

higher bitrate = lower noise floor
higher sampling rate = extended frequency response

Since 16/44.1 already has a noise floor so low you would have to turn the volume up to hearing damage levels to hear it all, and since it has a frequency response that covers the full spectrum of human hearing, what can you possibly expect to hear?

 

[KeithEmo]

That makes sense - sort of (except for the part where they neglect to mention it).
If they were honest, they could put a positive spin on it, and claim that "the CD layer is optimized for portable equipment" or some such similar claim.
I also suspect that sometimes it occurs for less planned out reasons (where the mixing engineer, knowing that the SACD master is for "the audiophile version", masters it differently according to whatever they believe "audiophiles" would want it to sound like.)
If people expect the DSD layer to sound different, in a certain way, and will interpret that as "better", it only makes sense to give them what they want...
Of course, sadly, it also fuels the claim the "SACDs sound better than CDs" since many people are under the mistaken assumption that they are making a fair comparison when they switch back and forth on a given disc.

Interesting that it is a different mix, however it is not uncommon for the CD layer of a SACD to be more compressed. The rationale of the labels is that the CD layer is better for everyday listening (eg in noisier environments) while the more dynamically mastered SACD layer is for quiet listening rooms.

 

[bigshot]

However, like many people, I do occasionally turn up the volume on the quiet parts to hear something better.

That's the textbook definition of a bad mix. At normal listening volume, everything should be clear. You shouldn't have to reach for the volume knob when you listen to music. You should just listen to music without being interrupted like that. Imagine if I went to the Disney Concert Hall to hear the L A Philharmonic and I got up out of my seat and walked up on the stage because I wanted to hear the harp. They'd kick me out and rightly so. Stay in your chair! Don't touch that dial!

There's a proper way to appreciate music. You wouldn't go to an art museum and pull out a magnifying glass to examine the cracks in the impasto. The same goes for music. You should appreciate it for the entire composition, not go out of your way to look for flaws. We are imperfect humans living in an imperfect world. There's no pleasure in focusing on the imperfections to the exclusion of the beauty around us. In fact, that is an endless rabbit hole that will drag you down into audiophool places you don't want to go. Listen to music, don't diddle with your volume knob!

 

[KeithEmo]

I clipped the quote because it was getting terribly long......

2)
I can't speak for other people's speakers or their capabilities.
I'm also not claiming that I choose my speakers based on their response above 20 kHz.
I'm simply disproving the claim that "speakers don't have useful response above 20 kHz".
(The claim is wrong because, clearly, at least SOME speakers do.)

3)
Offhand I have no idea who is filtering what where... nor what the response limitations are on the specific microphones they may have used (or their synthesizer's frequency response if it's electronic).
I simply prefer for my equipment to be able to reproduce the entire musical spectrum if it's there in the recording.
(I consider that to be the definition of "accurate reproduction".)

4a) - 5a)
I agree that most of us don't actually need to accurately reproduce CD-4 content....
Likewise, I agree that most of probably will never use the erase carrier tone leakage (and, yes, special tape heads and electronics were used)....
My point, which I stick by, is that we don't always know our future needs in advance....
(And, if someone says "if you turn the volume up at 14:27 you can hear the musician cough" or "a bat got into the studio and you can see his squeak on a spectrum display at 34:10" I would prefer to be able to try it.)
And, yes, I would prefer a full-spectrum copy of the tape rather than one that has been limited to "what I need" - based on someone else's ideas of what that comprises.

Again, I simply see the goal as an absolutely perfect and complete rendition of the original.
And I see ANY AND ALL LIMITATIONS as a compromise.
And I prefer to avoid compromises unless absolutely necessary (or unless I agree to them).

Someone asked, jokingly, whether videophiles would prefer it if their TV could reproduce gamma ray frequencies so accurately they could be burned by the video of a nuclear explosion.
While the example is absurd, I would say that the answer is technically yes.
We would be better off if our video display could reproduce frequencies from DC to gamma rays - and then we or the producer of the video could DECIDE which ones to limit or omit.
(I'm sure the government would cheerfully add some sort of safety standard to cover that aspect of things.)
However, in that case, both the safety and technical limitations JUSTIFY setting a standard that falls short of that lofty goal.

As a counter-example I would offer the color purple.
Most NTSC-standard TV sets don't reproduce a deep yet bright saturated purple color very well at all.
Therefore, most SD videos rarely show anything in bright purple - like a shiny amethyst necklace.
We have a situation where "the equipment can't play the signal" and "there's no point in including the signal because nobody's TV will be able to play it".
It has reached a point where set designers avoid using certain colors because they know those colors will be poorly reproduced.
(And, when you play those few videos that ignore the limitation on a full-spectrum monitor, you see an immediate difference. Now, interestingly, delivering the full color gamut - or more of it - is a major selling point of 4k HDR.)

6)
I do have an easy question for you......
If "quality assurance is admirable" (your words), then why are you arguing against it?

7a)
I'm not concerned with fakes... or of how to distinguish them from legitimate copies.
In either case, a properly provenanced legitimate original that is undamaged will virtually always be worth more than one that is damaged.
Likewise, even an undamaged copy will be worth more than a damaged copy.
Notice that you said "VISIBLY" repaired..... while I didn't include it......
While visible repairs are surely worse than invisible ones.... damage is still damage.
If someone buys your expensive vase, and finds a repaired crack when they x-ray it, they probably WILL sue you if you claimed it hadn't been repaired.
"Visible" and "nonexistent" are not the same thing at all...... (otherwise a perfect forgery of a Rembrandt would really be just as good as the original).

Yes, if you can find a lossy compression method whose output will be INDISTINGUISHABLE from the original, using our ears, or any test we can device, then it will be perfect.
(Except, of course, by definition, it WON'T be lossy at that point.)

The fact that some people find it easier to identify imperfect originals than perfect copies is irrelevant....

I am WELL aware of how lossy compression works.......
And it ALL amounts to "discarding content that someone else has decided I won't notice is missing".
(And I have very little faith in the choices made by other people - especially when those choices are often made based on "what 95% of people won't notice" rather than specifically on "what ***I*** won't notice".)

8)
You're entirely incorrect in one regard......
While the compression level used in JPG is variable - THERE IS NO SETTING IN THE JPG STANDARD THAT EXACTLY REPRODUCES THE ORIGINAL.
Even if you set a compression level that results in a file that is larger than the RAW file it is still lossy (you cannot retrieve the original pixels exactly).
A RAW file contains all of the information that was available from the camera; which is why it gives you the most flexibility and retains the most information.
An "uncompressed TIFF file" contains most but not all of the information; but is a much more standard format - which justifies the slight loss.
A JPG file discards a significant amount of information.... because a lot of information is approximated or discarded outright.

I also suspect you haven't edited many image files.... especially JPGs.
The method of compression used by JPG is applied to square zones of the image (so each square of a certain size is processed separately).
The compression is applied with certain constraints that ensure that the seams between adjacent squares won't be visible.
However, those constraints are based on several assumptions, including the conditions under which the image will be displayed, and the characteristics of the monitor or printer that will be used.
As a result, even though a given JPG may look "visually perfect" under very certain conditions, their failings tend to become unpleasantly obvious when you change the conditions.
Specifically, when you adjust the brightness, contrast, or color saturation, the sharp discontinuities at block boundaries become visible, and you get that "JPG blockiness artifact" that so many people find annoying.
(I'm ignoring the fact that virtually every pixel has been changed from its original color even though the net overall difference may be "imperceptable" to most people.)

Of course, there is an "original loss" because, while many cameras may exceed the performance of the human eye in specific ways, none so far exceeds the human eye in ALL regards simultaneously.

Take a picture of this posting, on a really sharp screen, and save it as RAW and as JPG.
Blow it up so you can see the individual letters.
I would be very surprised if you fail to see odd ghosts and echoes around the edges of the letters.
(Because JPG was optimized to compress "pictures with continuous tones", which works well for photographs, but does relatively poor job on sharp edges and narrow lines.)

And, yes, there is a parallel in audio.....

If the audio recording was originally recorded in MP3, then, yes, the most accurate rendition of the original would be a copy of that MP3.
And, if it was recorded at 24/96k, then that file would be the most accurate version.
(Of, assuming it was mixed, that would be the final output file that was sent to the production house.)

BUT, if you convert that MP3 file to 24/96k, you will be able to get back a very close approximation of the original MP3 file.
HOWEVER, if you convert that 24/96k file to an MP3 file, you will NOT be able to get back a close approximation of the original 24/96k file.

At all comes back to your original assertion that "none of the stuff we're throwing away really matters"...
I will agree that, for most people, on the equipment they'll be using, they probably won't notice the difference.....
But, if I was hoping to use the signature of that background noise to determine the brand of recorder the original was recorded on, then you have ruined my project if you discard it.
YOU have decided that the background noise is NOT "a meaningful part of the recording" - but perhaps we don't all agree.
And, yes, if you're the recording engineer, then, by definition, any such decision you've made is "correct".

9)
We seem to be in perfect agreement on the final conclusion.
Personally, I want the unlimited version, and am willing to pay a little extra for it.
However, I do absolutely agree that many people may find no benefit to high-resolution files, and shouldn't buy them.

Incidentally, as for your other comment...... ALL audio circuitry has bandwidth limitations.....

They may be inherent in the circuit design or specifically imposed using an external filter - but they are necessary.
Essentially ALL circuitry produces noise... and, in general, that noise has an infinite bandwidth.
So, even though your microphone may not pick up 30 kHz, the active devices in your microphone preamp are putting out SOME LEVEL of 30 kHz noise (and some level of 10 mHz noise).
If you're planning to digitize that output, you must filter out any significant noise above the Nyquist frequency to prevent audible aliasing and other errors.
You also MIGHT have issues, as some have suggested, with intermodulation distortion caused by high-frequency noise (or content).
And you really wouldn't want to die because your amplifier accurately amplifies the leaked noise it picks up from your microwave oven or cell phone.

There's also another technical issue that MOST (probably all) amplification circuits exhibit phase shift that increases as the frequency increases.
As a result, at some very high frequency, the phase shift is such that the circuit's negative feedback becomes positive feedback, and the circuit oscillates.
(And virtually all modern audio circuitry employs negative feedback.)
Because of this, audio circuits are designed so that their gain falls at high frequencies..... which is another way of describing "a bandwidth limiting filter".

With older equipment, this limitation was often more or less a random result of the overall design.... or simply a matter of luck.
However, with modern design, it is assumed that each piece of equipment has been designed to "protect itself from anything it would have a problem with".
Basically, with analog equipment, the circuit is designed so, as you look at higher and higher frequencies, the gain reaches unity or lower before the phase shift reaches a dangerous point.
(Most modern amplifiers have a frequency response that starts rolling off significantly around 80 kHz or so - as that is considered to be "well outside the audible frequency range".)

9. I have no issue with capturing the original in as complete a representation as possible, and distributing that version. I have big problems with the general assumption that high sampling rates and bit depths result in categorically better sound, and question strongly the value of ultrasonic content over 20kHz and the ability of the average or even high-end user to get that energy to his eardrums (much less to actually hear it). I do think there are cases where ultrasonic energy, real or distortion products, can cause problems in devices not capable of handling them without distortion. I do believe there are cases where the process of band-limiting audio by the use of certain devices results in intermodulation distortion that can be folded down into the audible region making wideband audio sound better for a reason other than the ultrasonic content. I also believe that those cases are fewer today than say 25 or 30 years ago, and the real solution is to test for high frequency intermodulation and deal with the cause rather than to band-aid a solution either by passing a wider bandwidth or limiting ultrasonics by filtering. Only a few decades back there were audio products that included ultrasonic filters to prevent those signals from wreaking havoc in other devices. Hopefully today those devices are few, but I doubt they're completely gone.

 

[castleofargh]

That makes sense - sort of (except for the part where they neglect to mention it).
If they were honest, they could put a positive spin on it, and claim that "the CD layer is optimized for portable equipment" or some such similar claim.
I also suspect that sometimes it occurs for less planned out reasons (where the mixing engineer, knowing that the SACD master is for "the audiophile version", masters it differently according to whatever they believe "audiophiles" would want it to sound like.)
If people expect the DSD layer to sound different, in a certain way, and will interpret that as "better", it only makes sense to give them what they want...
Of course, sadly, it also fuels the claim the "SACDs sound better than CDs" since many people are under the mistaken assumption that they are making a fair comparison when they switch back and forth on a given disc.

yup, and as is the case most of the time, instead of simply advocating for the good aspect of what they do, we end up with no information at all so that people can mistake mastering for format sound. or worst, we get the kind of BS marketing/propaganda we got for Daft Punk's Random Access Memory. where they went to interview the various guys involved and have the engineer say the dumbest most dishonest stuff about the reason why they made different masters for the different media.
the guy (forgot his name, the french dude, not the US guy doing the recording), implied that some stuff wouldn't "translate" on CD so they did it differently... I could imagine him talking to his friends in the industry in one of those "and then I told them" meme

it gets even "funnier" when you actually listen to the versions. wow! yeah, now I get it, no way that massive low end boost could have worked on the CD version, CDs can't handle a bass boost, everybody knows that. it's just not dynamic enough...

it's infuriating because they take us for the dumbest milk cows ever. but what makes it worst is that they never needed to lie. they could have just said, "when you buy the SACD you get several versions of the song as an exclusive". or indeed argue that for listening in a noisy environment, the more compressed version would be more practical. there are numerous ways to sell it as a good thing. but at some point, marketing is just so used to lie about made up benefits of high res, that they forget how some products could actually sell themselves for what they are.

 

[KeithEmo]

I kind of agree.....

Except that you're mixing philosophy and technology.
Whether you SHOULD diddle with that knob is a matter of philosophy; what happens when and if you do is a matter of technology.

And, saying that somebody shouldn't care about something is philosophy...
While saying that it isn't there is a matter of fact.
(There is a one-way philosophical overlap that says that worrying about something that is "meaningless".)

I should note that you seem to have highlighted my main issue with this thread.....
A lot of well meaning people have suggested that I shouldn't CARE about certain things.
And, that being a philosophical argument, I happily concede that they may in fact be right.

However, rather than convince me that it isn't important, they seem determined to insist that "it really isn't there so I'm silly to even be talking about it".
I'm sure you're right about their not being pleased if you decide to go walkabout during a concert.
However, I wouldn't suggest you were wrong if you said you wanted to try different seats so you could decide which one you want to purchase next time.
And I especially wouldn't try to convince you that, "since it is a professionally designed concert hall, all of the seats are good, so it doesn't matter".

And, while I would also cheerfully agree that many "audiophile claims" really are simply untrue....
I don't think that's true in this case (more accurately, while I also concede that it MIGHT be true, I don't think it's been PROVEN beyond reasonable doubt).

If I, as a scientist, say "'there's no audible difference", and ONE GUY shows up who can reliably hear one, then I've been proven both a liar and a fool (and, worse, a bad scientist).
I'd rather stick with "most" and "probably" - and get back to the PHILOSOPHICAL discussion of whether a difference that less than 1% of us can hear is worth bothering with or not.

That's the textbook definition of a bad mix. At normal listening volume, everything should be clear. You shouldn't have to reach for the volume knob when you listen to music. You should just listen to music without being interrupted like that. Imagine if I went to the Disney Concert Hall to hear the L A Philharmonic and I got up out of my seat and walked up on the stage because I wanted to hear the harp. They'd kick me out and rightly so. Stay in your chair! Don't touch that dial!

There's a proper way to appreciate music. You wouldn't go to an art museum and pull out a magnifying glass to examine the cracks in the impasto. The same goes for music. You should appreciate it for the entire composition, not go out of your way to look for flaws. We are imperfect humans living in an imperfect world. There's no pleasure in focusing on the imperfections to the exclusion of the beauty around us. In fact, that is an endless rabbit hole that will drag you down into audiophool places you don't want to go. Listen to music, don't diddle with your volume knob!

 

[KeithEmo]

Perhaps they just want to force everyone to buy ALL the versions .

I have a similar issue with certain symphonic metal bands I like.
They are very international, and each of their albums is published in a variety of different "country versions".
Annoyingly, each has one or two bonus tracks, and each national version has DIFFERENT bonus tracks.
They insist that it's "to please different fans in different countries"....
But it seems clear to me that they simply want to force their serious fans, who don't want to miss any songs, to buy a copy of EACH version to get all the bonus tracks.

(One band, as I recall, Edenbridge, actually had the courtesy to issue a separate album of the various bonus tracks so you could buy them all at once.)

yup, and as is the case most of the time, instead of simply advocating for the good aspect of what they do, we end up with no information at all so that people can mistake mastering for format sound. or worst, we get the kind of BS marketing/propaganda we got for Daft Punk's Random Access Memory. where they went to interview the various guys involved and have the engineer say the dumbest most dishonest stuff about the reason why they made different masters for the different media.
the guy (forgot his name, the french dude, not the US guy doing the recording), implied that some stuff wouldn't "translate" on CD so they did it differently... I could imagine him talking to his friends in the industry in one of those "and then I told them" meme

it gets even "funnier" when you actually listen to the versions. wow! yeah, now I get it, no way that massive low end boost could have worked on the CD version, CDs can't handle a bass boost, everybody knows that. it's just not dynamic enough...

it's infuriating because they take us for the dumbest milk cows ever. but what makes it worst is that they never needed to lie. they could have just said, "when you buy the SACD you get several versions of the song as an exclusive". or indeed argue that for listening in a noisy environment, the more compressed version would be more practical. there are numerous ways to sell it as a good thing. but at some point, marketing is just so used to lie about made up benefits of high res, that they forget how some products could actually sell themselves for what they are.

 

[bigshot]

As I said to the other guy... My grandmother could hear the noise floor of 16 bit if you turn the volume up high enough. I'm sure you could hear the noise floor of an SACD too if you crank the volume. That isn't any surprise. If I grab a magnifying glass I can see pixels or film grain or halftone dots in photographs. If I look close enough I can see a single grain of sand on a beach. Heck, with a scanning electron microscope I can see an atom!

Turning up the volume of a CD so high that you can hear the noise floor is totally irrelevant to the purpose it was designed for. The purpose of a CD is to put it on your stereo and listen to music in your living room. Expecting 16/44.1 to wash your dishes for you or mow the lawn or split atoms is ludicrous. It was never designed to do that. Who cares? Why are we even talking about it? Beats me.

When an argument goes this far off the rails, you really don't have an argument any more. I see a lot of arguments around here lately that are seriously out of touch with reality. I think there is some sort of virus going around or something!

We're in a forum about home audio. For the purposes of listening to music in the home, 16/44.1 is all you need. Properly designed DACs (even cheap ones) are audibly transparent and do the job too. It's fine to talk about moonbeams and unicorns, but it's more entertaining to express those thoughts in small doses.