I Wish SACD & DSD Would Go Away

QuadraphonicQuad

Help Support QuadraphonicQuad:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
One thing I find intriguing about this discussion is the notion that you have to drop a billion dollars on a system to play SACD or DSD properly.

'Intriguing' would a polite way of putting it. 'Laughable' is closer to the mark for me.

I have a very modest system and SACD and DSD sound great, regardless of what my player or receiver are actually doing.

On mine, a great recording sounds great, regardless of format.
 
Me....I've been in audio forever and have always tried to keep up with the latest trends but even that's become nigh impossible. I have NO plans for Dolby Atmos [being a full range speaker man most of my life] and if I can ever get my main listening room reconfigured, my last statement will be buying these uber whacky expensive Meridian self powered speakers with built in MQA. No DSD downloads IN my future as the system will be PCM ALL the way [no digital to analogue conversion necessary].

That would truly be a magical system.
 
I don't. Maybe it's buying into blu-ray and some of the fantastic sounding audio on the medium, maybe it's the speakers, or the current receiver not decoding DTS as nicely as the previous one, I don't know, but there's a sound (or lack of sound, if you like) that seems to accompany the data compression when I listen.

Or maybe a cognitive bias.
 
When one ponders the alternative to DSD/SACD namely lossy codecs like DTS and Dolby Digital, thank the stars we have high resolution formats like DSD/SACD, DVD~A and BD~A/V to remind us ALL how VERY SPECIAL high res formats are!

For those who are 'content' with DTS [and yes, it's better than DD], sorry you really cannot hear the difference between that and higher res formats.

It is system dependent. That's for damn sure.

Far more likely, it is sighted-listening dependent.
 
I recall being blown away by the sound of my Millennium 2.4.6 DTS Decoder/Preamp, in patricular the DTS CD "BILLIE MYERS - Growing Pains". DTS have the audacity to say that they use compression to improve the sound!

That seems odd, if they mean lossy perceptual data compression. Can you point me to DTS's literature claiming this?

Actually there is some truth in that statement as for one thing DTS CD's are 20bit where as CD's are only 16bit. I've been a bit less impressed with the sound of DTS from my universal DVD players, I don't know if the built in decoders are of lesser quality than that of the Millennium or if it's just that it's up against the sound of DVD-Audio and SACD. The difference between DTS and Dolby however is like night and day!

:rolleyes:
 
Or maybe a cognitive bias.

I don't think it is. For clarity, I'm talking about DTS CD rather than the system found on DVD. I'm not particularly snobby about Dolby Digital, and I'm not particularly fond of DSD, which I thought had a sonic signature, despite the hype leading cognitive bias in the other direction.
 
One thing I find intriguing about this discussion is the notion that you have to drop a billion dollars on a system to play SACD or DSD properly.
I have a very modest system and SACD and DSD sound great, regardless of what my player or receiver are actually doing.

Amen. People are always looking for problems.
 
Absolutely. If you spend 1000x more on a system you won't get one any where near a thousand times better.

One thing I find intriguing about this discussion is the notion that you have to drop a billion dollars on a system to play SACD or DSD properly.
I have a very modest system and SACD and DSD sound great, regardless of what my player or receiver are actually doing.

Amen. People are always looking for problems.
 
That seems odd, if they mean lossy perceptual data compression. Can you point me to DTS's literature claiming this?



:rolleyes:

It was from a white paper about DTS titled "An Overview of the Coherent Acoustics Coding System" here is one quote.

"The DTS Coherent Acoustics audio compression algorithm was designed with the primary objective of
significantly improving the quality of audio reproduction in the home, beyond that of conventional compact discs"

http://www.mp3-tech.org/programmer/docs/dts_whitepaper.pdf
 
It was from a white paper about DTS titled "An Overview of the Coherent Acoustics Coding System" here is one quote.

"The DTS Coherent Acoustics audio compression algorithm was designed with the primary objective of
significantly improving the quality of audio reproduction in the home, beyond that of conventional compact discs"

http://www.mp3-tech.org/programmer/docs/dts_whitepaper.pdf

And followed specifically by:
"Consumers would benefit from more accurate sound recordings that utilized a wider range of audiofrequencies, played back through more loudspeakers"

As relates to CD, half of that is tendentious bosh (the 'wider range of audio frequencies' stuff), half isn't (certainly, home audio can be improved by going from 2 channel -- CD -- to multi. But that can be achieved without lossy compression. ) And even if 'wider frequency' beyond 22kHz mattered, there is nothing about DTS's *compression algorithm* per se that would make it sound better. The 'better' would come from simply using wideband digital files as input and permitting ultrasonic frequencies at output (compressed or not). In 1999 that was already happening with CD production (96K masters), and lossless DVDA and SACD , whose output is also wideband, were just around the corner as consumer formats. And as we know, the 'dramatic improvements in reproduction quality' resulting from high SR playback has never been scientifically documented for any.

It's also nonsensical as presented, since the white paper immediately goes on to argue that 'linear PCM' (of which CD is one format)is inferior because it is *inefficient* -- it over-captures frequencies and amplitudes that music tends not to display and humans tend not to hear! That's a fine argument for perceptual data compression, but it contradicts the earlier claim and it is ludicrous as an argument for *superior accuracy*. At best it will be transparent to 'linear PCM', not *audibly better*
And that is what they got for results, when they did blind tests (reported, with no data, later in the paper). They did not, btw, compare CD rate to hi-rez audio.

In short, don't necessarily believe everything you read in a company white paper.
 
People usually don't. But it usually is. That's a characteristic of 'sighted' audio format bias.

I get that, I am quite a skeptic with some of the stuff that is claimed. I changed parts of my system and noticed change. Maybe because I've been exposed to more hi-res. I don't think it's unreasonable for a 6 (or even 8) channel DTS signal compressed to fit on a CD to have a difference in sound. It seems you will accuse me of bias for that assumption, so good for you. I'm fully aware of sighted audio format bias. You didn't address my comment about DSD sonic signature despite sighted bias and hype suggesting it should sound brilliant. If I could do a null test on DTS CD vs a hi-res I'd be very open to do so. But how about you get off your high horse and realise there is the physics to back up that it could be audibly different. Special polish or a green pen only requires a null test. I don't believe they make a difference and nobody appears willing to supply to simple evidence to back up that they do, so lay off the format bias with me. Maybe it coincides that they're not great sounding discs, maybe the decoding in the new system isn't as good. I do have some knowledge, I do have some experience and you could do with being less of a jerk about it.
 
At best it will be transparent to 'linear PCM', not *audibly better*
And that is what they got for results, when they did blind tests (reported, with no data, later in the paper). They did not, btw, compare CD rate to hi-rez audio.

On the one hand you accuse me of sighted audio bias, on the other hand you say this. That seems contradictory.
 
I get that, I am quite a skeptic with some of the stuff that is claimed. I changed parts of my system and noticed change. Maybe because I've been exposed to more hi-res. I don't think it's unreasonable for a 6 (or even 8) channel DTS signal compressed to fit on a CD to have a difference in sound. It seems you will accuse me of bias for that assumption, so good for you. I'm fully aware of sighted audio format bias. You didn't address my comment about DSD sonic signature despite sighted bias and hype suggesting it should sound brilliant. If I could do a null test on DTS CD vs a hi-res I'd be very open to do so. But how about you get off your high horse and realise there is the physics to back up that it could be audibly different. Special polish or a green pen only requires a null test. I don't believe they make a difference and nobody appears willing to supply to simple evidence to back up that they do, so lay off the format bias with me.

It bemuses me when someone says 'I get that' re : perceptual bias, then immediately goes on to assert differences from sighted listening. (I changed parts of my system and noticed change....DSD sonic signature....)

The fact that a multichannel signal is 'compressed to fit on a CD' isn't, itself, an argument that it is likely to sound different; lossy perceptual compression algorithms matter...and they work. The input and output signals surely won't 'null'...but that's a feature, not a bug. By design, masked and predicted-to-be inaudible parts of the signal are removed. Absent psychoacoustic data that predicts we *should* hear a difference when the DTS algorithm is appied-- and not just a tiny one, but one obvious to then likes of you even on casual listening -- and absent personal testing data, your argument is from incredulity.


Maybe it coincides that they're not great sounding discs, maybe the decoding in the new system isn't as good.

In which case, it's not DTS vs anything.

I do have some knowledge, I do have some experience and you could do with being less of a jerk about it.

If you have better evidence, by all means offer it.

If you really do 'get it', you would accept the 'accusation' of sighted bias as *at least as if not more* plausible an explanation for what you perceived, as the explanations you offer.
 
I don't see how.

And btw you aren't 'accused of' sighted bias. It's not a crime. It's a natural characteristic of human perception.

You say on the one hand that I'm not hearing any difference due to bias then you say that DTS can't possibly sound better than CD, at best transparent, so at best they would sound perceptably the same. At best. Your words. So you argue two sides of the same debate without looking at my reasoning or taking into account any knowledge or experience or skeptism I have which appears to display a bias of your own.
 
The fact that a multichannel signal is 'compressed to fit on a CD' isn't, itself, an argument that it is likely to sound different; lossy perceptual compression algorithms matter...and they work.

I get that. Also the fact that it is compressed is an argument that it *can* (not necessarily does, but can) sound different. I don't think I expected it to sound worse, I didn't want it to sound worse - what gain is that to me? You refute sighted bias, and then you refute that I've not liked the sound that according to sighted bias should sound superior. Nobody told me about DSD sonic signature (or maybe converter sonic signature), I wasn't expecting such a thing, I wasn't directed to such a thing. I didn't even know about the noise shaping: I had no reason to believe or expect something like that in the make up of DSD. You make arguments for both sides of the debate and you refute both sides of the debate. Audio data compression is clever, I know that. Losing a two thirds of the data on lossy compression to make something quite listenable is clever. Minidisc was incredible, I have no knowledge of the compression algorithms used for that, nor did I use it at home beyond a some all in one system. I liked it though.
 
You say on the one hand that I'm not hearing any difference due to bias then you say that DTS can't possibly sound better than CD, at best transparent, so at best they would sound perceptably the same. At best. Your words.

Still not getting the contradiction here -- anyone?

Re: DTS claims, what I said was that their claim for DTS sounding 'better than' CD is based on two ideas. 1) The dubious idea consumers are readily perceiving benefits of higher sample rates and 24 bits (separate from whatever other mastering differences there are), and 2) the well-supported idea that more channels means more realism (or at the very least, an extremely easy-to-hear difference). Their claim #1 is not really that DTS compression itself makes their stuff sound better than CD, it's that they are compressing hi rez sources and maintaining (some of) that expanded bandwidth and bit depth through output. They don't even test the proposition re: CD, they simply perform a silly jujitsu move and start talking about the *inefficiency* of uncompressed PCM versus compressed data as a storage and delivery format, which is undeniable but not germane to the 'audible superiority' claim.

The only blind(?) listening test results they offer in that white paper were comparing input 'hi rez' (48kHz or more SR, 24bit) input, presumably multichannel, to DTS output, and they claim the result was *transparency*. Not 'better' (which would also mean 'different', aka, nontransparent). There's nothing in there explicitly comparing CD to hi rez, much less lossy compressed. Their information-sparse report is is follows:

Evaluating the
Coherent Acoustics system took place over a period of more than two years, and involved expert listeners
from the music and entertainment industry, using custom-built digital audio evaluation equipment installed in
professional listening rooms (figure 23). Using digital audio signals sampled at 48 kHz with a resolution of up
to 24 bits, these tests consistently demonstrated the transparency of Coherent Acoustics. Irrespective of the
audio material, there was no perceived difference between the original linear PCM signal and the coded
Coherent Acoustics version. More recent tests have also demonstrated the subjective transparency of
Coherent Acoustics at a sampling rate of 96 kHz .



This is the second time I'm explaining this, I hope it's clear now.
 
I get that. Also the fact that it is compressed is an argument that it *can* (not necessarily does, but can) sound different. I don't think I expected it to sound worse, I didn't want it to sound worse - what gain is that to me? You refute sighted bias, and then you refute that I've not liked the sound that according to sighted bias should sound superior. Nobody told me about DSD sonic signature (or maybe converter sonic signature), I wasn't expecting such a thing, I wasn't directed to such a thing. I didn't even know about the noise shaping: I had no reason to believe or expect something like that in the make up of DSD. You make arguments for both sides of the debate and you refute both sides of the debate. Audio data compression is clever, I know that. Losing a two thirds of the data on lossy compression to make something quite listenable is clever. Minidisc was incredible, I have no knowledge of the compression algorithms used for that, nor did I use it at home beyond a some all in one system. I liked it though.

I never argued that lossy compression *can't* make a signal sound different, that's a straw horse argument you've hopped on. Nor do DTS claim that, either.

I do say that to argue convincenly that it *does* make an audible difference in any particular case (and typically it's not claimed to be a subtle difference), requires more evidence that 'I heard it' -- either some evidence that the compression itself is highly compromised (e.g., using an mp3 codec at low bitrate), which is something *predicted* to sound different on psychoacoustic principles, or evidence from blind listening tests.

And that goes double for claims about DSD 'sonic signature' , and converters.

Face it. The default 'argument' from some here is that if A and B sounded different it's the format. It's just unquestioned. I always question it, and science backs me up on that stance. The looming elephant that is certainly in each of our rooms is hardly ever even acknowledged here. In fact to you it's an 'accusation', which speaks loudly of the audiophile mindset.
 
I see a lot of arguments that seem to focus on one or more variables. Often an argument that neglects to mention some element gets taken as a dismissal of it. It's good to keep an open mind and open ears. Some things are appropriate to dismiss. Some are worth paying attention to!

A lot of technically imperfect audio storage and delivery systems work at a high enough level that unless you have a very professionally tuned listening space and reference quality gear, you'll never be the wiser. That's a good thing right? And even if said elements of loss ARE perceivable on a well tuned system, it might be one of those 2% discussions. So we're usually in a really good place nowadays for hearing music as intended by the artist.

One of the points I try to make is that the mix quality and mastering quality are still the big factor. That's not a dismissal of any format discussion. Conversely, I'll suggest that since we HAVE lossless HD consumer delivery formats now that they should be used 'just because you can'. Might as well eliminate any format container variables if possible right?*

I've heard various lossy formats result in significant loss with telltale sonic signature. I've heard lowly mp3's of some things where I would have been none the wiser. I've heard excellent mixes really destroyed by 'portable device' style mastering (heavy limiting and then boosting and with heavy high end eq boosting) to the point that any loss from even the lowest quality formats (not just mp3... like poorly reproduced 8-track tape cartridges even) would be a moot point.

When I hear someone critique things like reducing 24 bit program to 16 bit, I'm with you in that I've heard many examples too. I think it should be avoided unless catering to older equipment that requires it. But then when someone talks about this in a way that suggests they hear loss more along the lines of what that poorly reproduced 8-track cartridge might do, I have to suggest that this isn't the variable that's the main cause of trouble in the scenario. Sample rate conversion is even more subtle than that generally. There might be some ringer recordings that are strikingly more susceptible to this than others but we're still having one of those top 5% discussions here. If you hear something more worthy of a 90% discussion for example, it's almost got to be a different variable at play than what you thought.


* I'm not critiquing every HD recording that comes along and having a shootout between 24/96 PCM vs. 24/192 PCM vs. DSD and thinking I'm hearing things. It's actually the other way around honestly! Damaged recordings are what tend to make further generation loss perceptible and sometimes to a surprising extent. The budget or consumer level AD & DA converters that run cleaner at HD vs. SD. 16 bit reductions. These are usually still pretty high level containers for audio but you get the odd case of hearing some loss. Starting with introduction of 24/96 PCM, I haven't heard a single example of any perceptible generation loss due to the container format. Even badly damaged recordings that tend to crumble into dust at the slightest touch are preserved intact. High fidelity recordings in DSD format that I've heard as well as DSD to 24/88.2 PCM conversions of them suggest to me that DSD is every bit the full quality container HD PCM is. If I hear damage, the root cause is always something else.

Following that, my suggestion is to use this format to avoid any "container loss" just because you can and then just set up the best monitor system that is reasonable to invest in for you. Further, when you hear something off, look at the bigger variables first.

My critique of DSD is simply that we didn't need a different digital "language" and to make the disc storage format that uses it (SACD) spin the opposite direction in the drive. None of that is aimed at fidelity improvement. It was a greedy power play move to get consumers to buy into a different system that would make it difficult to play anything from "Company A". Now if you want access to both, you need to invest in more equipment and/or learn some tech skills to deal with it which translates to less music and/or lower quality reproduction for many people. Thus I think it deserves to be ostracized.

Hear something funny?
When I do, I go in this order:
Is it just level differences between the two?
Is it the same mix?
Is it a mastering issue? (Or was one obviously mastered for portable device listening?)
Is one in a grossly lossy format? (mp3, lossy dolby, core dts)

Loss from sample rate reduction to SD or 24 bit reduced to 16 (not including snowballing degradation of already challenged recordings) is more of a comment like: What did you do that for when we have lossless 24/96 containers? I don't see 16/44.1 vs. 24/96 as going from 0 to 100. More like closing the gap from the high 90's to 100%.

When someone then comes along and suggests "Well, my speakers are only such and such so this lower setting is good enough." Spending money is spending money and choices need to be made. But when it turns into stubbornness with a switch setting... come on now! Now who's more interested in arguing vs. actually listening to some music?

When purchasing music, I see SD and lossy formats as a calling card for portable device style mastering and avoid them for that processing work which eclipses pretty much all discussions on container quality. Whereas HD formats are usually a calling card for well handled and presented recordings. There are exceptions in both directions of course but that's the general pattern. (I'm sure there's some mp3 out there that sounds better than some bluray disc release of the same for a ringer example. Which would be an example of a mastering disaster for said bluray.)


Well, have fun with that! :)
 
Back
Top