The sound quality of DSD?

[Pluto]

DSD is a pretty silly format if you think about it

Indeed. I am yet to encounter a DSD recording in which 70% of the data is anything other than (largely) random i.e. noise. Now I am no mathematician but it strikes me as rather excessive that you need that much noise content to act as a rather aristocratic form of dither.

The part about wondering why not use really good multibit converters...maybe they weren't as good in some specifications at the time?

At the time DSD converters started to appear (1997-ish) the record business had one dominating preoccupation: how to store its rapidly deteriorating IP in a format that was an order of magnitude better than that being offered to punters (16/44). At that time, some 20 bit PCM converters had started to appear and even the occasional 24 bit unit, but tests revealed that if 18 bit accuracy was achieved you were operating at the then state of the art.

So DSD appeared to be the answer to the industry's prayers. It was intended as an archive format and nobody (seriously) thought that it might become a production option. But the engineers who devised DSD hadn't reckoned with the wily marketing droids who were looking at options for the “next generation” beyond Red Book CD that absolutely, beyond any shadow of a doubt, had to have the means of being locked up so tight that a repeat of the CD-R copying farce could never happen again.

And one of the most useful tools in achieving that degree of lock-up was the use of a format so obscurely different from PCM that it was like trying to handle hot coals…

 

[JoachimStrobel]

SACDs survive, or even grow, because the format offers the highest copy protection security, so it seems. Sure, some Oppos can crack a SACD, but there is no SACD drive for a PC which secures it from the endlessly hacking community, unlike DVD and Blu-Ray. Mch PCM via HDMI seems already to be guarded like nothing else, Mch DSD tops even that. Talking people into believing DSD is the best is the best thing the industry can do for copy protection. The ease of Red-Book copies must have left deep wounds...

 

[firedog]

Most of your arguments rejecting DSD are irrelevant. Because you don't get it. Your techical arguments (some of which are BS) don't matter: People like DSD because it sounds different and and also better to them. That's their taste in sound. Doesn't matter why, and doesn't matter if you "approve" of the technical aspects of DSD or not. It also doesn't matter if you like it or can hear the difference.
The copy protection argument doesn't hold water anymore either. DSD downloads are widely available for people who like the format. Yes, most of those are classical music - but so is most of the market for DSD and SACD.

 

[Jimbob54]

Most of your arguments rejecting DSD are irrelevant. Because you don't get it. Your techical arguments (some of which are BS) don't matter: People like DSD because it sounds different and and also better to them. That's their taste in sound. Doesn't matter why, and doesn't matter if you "approve" of the technical aspects of DSD or not. It also doesn't matter if you like it or can hear the difference.
The copy protection argument doesn't hold water anymore either. DSD downloads are widely available for people who like the format. Yes, most of those are classical music - but so is most of the market for DSD and SACD.

When you say "sounds different" you're making no subjective judgement. So like kit with distortion by design, some may like, others not? All good so far.

But isnt the spiel that its more accurate to the source ? So if it sounds different (to PCM) is that because it is more or less fidelius to the source?

 

[Head_Unit]

copy protection

The copy protection argument doesn't hold water anymore

Actually the copy protection was a boon but also a big impediment. To hear your disc they had to STOP the stamping plant, press your physical disc, then resume production. You couldn't just dash off a CD-R for your client to listen to in their car on the way home. That was a big crimp that a lot of production-side folks complained about when I'd go to those surround conferences they used to hold. Nowadays I'm sure that's been worked around.

 

[firedog]

When you say "sounds different" you're making no subjective judgement. So like kit with distortion by design, some may like, others not? All good so far.

But isnt the spiel that its more accurate to the source ? So if it sounds different (to PCM) is that because it is more or less fidelius to the source?

Don't know and don't care. I don't really think that has much to do with it's appeal. The people that like it don't really talk about that. They talk about how they prefer it's sound.
It's produced differently so it sounds different. Play around in a program like Roon or HQP and you can hear the difference. Either you like it or you don't.
Do you really think a recording done in DSD is somehow "innaccurate"? Obviously it isn't. It's a different format than PCM and has some advantages and some disadvantages.

 

[RayDunzl]

Especially when the digital volume is reduced with DSD native,

Your DAC can digitally attenuate DSD?

 

[Buster Chestnut]

Most of your arguments rejecting DSD are irrelevant. Because you don't get it. Your techical arguments (some of which are BS) don't matter: People like DSD because it sounds different and and also better to them. That's their taste in sound. Doesn't matter why, and doesn't matter if you "approve" of the technical aspects of DSD or not. It also doesn't matter if you like it or can hear the difference.
The copy protection argument doesn't hold water anymore either. DSD downloads are widely available for people who like the format. Yes, most of those are classical music - but so is most of the market for DSD and SACD.

I never knew there was so much dislike for DSD before this thread; it’s been entertaining

 

[firedog]

I never knew there was so much dislike for DSD before this thread; it’s been entertaining

It's unfortunate that some of the posters on this forum seem to participate mostly to put others down, and seem to think if someone has a different taste/preference than them, or reacts differently to a sound, that that is illegitimate. Somehow it is hard for them to understand that even if we listen to the same thing, we may not hear the same thing. There are personal differences.
Example: my bright and harsh sounding can be your accurate and detailed. And so on.

 

[theglobalelite]

DSD = PCM

A highly oversampled, low bit/sample and noise shaped PCM. I wouldn't even bother giving it a second thought... There is a reason DSD is irrelevant everywhere aside from the likes of audiophiles that generally don't know any better.

Hey now, if you like surround mixes of albums DSDs/SACDs are really the only game for a lot of that. So it's not completely useless.

 

[firedog]

Hey now, if you like surround mixes of albums DSDs/SACDs are really the only game for a lot of that. So it's not completely useless.

Yep. One of the markets for SACD is mulitchannel classical. In that market, that's what the buyers want, and they don't even care if the master is PCM.

 

[Frank Dernie]

The quality of the recording and mastering make a much bigger difference to SQ than the format it is distributed on IME.

 

[SIY]

/looks for actual evidence that there is any sound difference with the same master

/doesn't see that mentioned anywhere

 

[bennetng]

Based on listening to the sound quality of DSD native from Cirrus Logic chips like that in Sonata HD Pro I find:
• S/N ratio of DSD at high volume is bad. At high volumes, you can hear noise. Not good at all!

Your DAC can digitally attenuate DSD?

Yes if not bypassing the DSD filter and multibit DSM. In other words, neither "native" nor "direct". Only the DSD sample rate is retained.

CS43131

CS4398

One should realize that DSD is a marketing term. DSD is a form of DSM, but DSM is not DSD. Just like apple is fruit does not mean fruit is apple. DACs accept 1-bit DSD and multibit PCM, only these two formats, regardless of what happens outside of the DAC (e.g. DXD, DSD "wide" and such).

The recent AKM chipset is a good example of this topology. The AK4191 DSM chip works in 64-bit, but accepts DSD and PCM input, and outputs 7-bit DSM. AKM doesn't call it "multibit DSD". Then the output is being sent to a proprietary 6-bit format to the AK4498 DAC.

DSD as a "native" format is only relevant to boutique DACs with custom implementations or ~30 years old converters with real 1-bit DSM and 64fs oversampling like CS4328, ironically the whole datasheet does not contain the word "DSD".

 

[polmuaddib]

Yep. One of the markets for SACD is mulitchannel classical. In that market, that's what the buyers want, and they don't even care if the master is PCM.

I am not so sure. Ther is an old sa-cd.net site which is still active and they bothered (when possible) to document what master was. Were it analog, PCM or pure DSD. In the comment section a lot of people commented and cared about the sound, recording.
Of course, pure DSD recording was always and still prefered. And i believe that it sounds better, but likely because there were minimal production processing and they relied more on proper recording techniques.
If there is any difference, it would be because of production.
And when you stop to think about it, pure DSD chain (not PCM or analog tape masters converted to DSD) might have a value just because you can't process the life out of recording, like you can with PCM. Maybe I am wrong, but that was always the main advantage and disadvantage of DSD - no digital processing, or at least minimal processing. I know there is Pyramix DAW and what not....

 

[Pluto]

The quality of the recording and mastering make a much bigger difference to SQ than the format it is distributed on

An odd irony is that there is no industry-agreed alignment of levels (for either recording or playback) between DSD and PCM. So unless you take steps to manage this (sometimes not easy) you can get a DSD stream derived from exactly the same source as corresponding PCM and there might be a few dB between them. Furthermore, this difference might vary between devices.

And we all know what happens when comparing identical sources at differing levels…

 

[Feanor]

The DSD vs. PCM issues are beyond my comprehension, and I have no technical opinion on the subject.

... Meanwhile I'm fine with well-recorded and well-mastered 16 bit / 44.1 kHz, (PCM). The major issue with sound quality today is the human aspect of recording & mastering, NOT the digital medium for either production or distribution.

 

[Kal Rubinson]

ep. One of the markets for SACD is mulitchannel classical. In that market, that's what the buyers want, and they don't even care if the master is PCM.

Sure, we do. Any release not in the same data format as the source should be investigated for provenance and process.

 

[Frank Dernie]

An odd irony is that there is no industry-agreed alignment of levels (for either recording or playback) between DSD and PCM. So unless you take steps to manage this (sometimes not easy) you can get a DSD stream derived from exactly the same source as corresponding PCM and there might be a few dB between them. Furthermore, this difference might vary between devices.

And we all know what happens when comparing identical sources at differing levels…

A friend sent me some LP rips he had done in DSD and PCM to compare the DSD files were a bit louder and, surprise, surprise sounded a bit clearer and mor dynamic.

 

[mansr]

Indeed. I am yet to encounter a DSD recording in which 70% of the data is anything other than (largely) random i.e. noise. Now I am no mathematician but it strikes me as rather excessive that you need that much noise content to act as a rather aristocratic form of dither.

As an intermediate format inside a DAC or ADC, it makes sense. Traditional multi-bit DAC designs (resistor ladders and others) don't scale well to increasing bit depths. Even at 16 bits, the component matching has to be extremely tight to avoid non-linear distortion, and they require tricky de-glitching circuitry as well (ask Schiit what happens when you ignore that). A high-rate 1-bit "DSD" signal, however, needs only a stable clock and a fast switch followed by a low-pass filter to produce an analogue output. By the early 90s, digital electronics operating at a few tens of MHz had become cheap, and this is when the "bitstream" DAC chips appeared. Now creating the 1-bit signal is not without challenges, and the chip manufacturers soon moved to multi-level sigma-delta designs since even having a mere three levels instead of two greatly simplifies some aspects. As long as the number of levels is reasonably small (less than ~100), the A/D stage remains fairly simple without the need for impossibly tight tolerances. At the same time, an increased number of levels reduces the amount of noise that has to be removed, thus simplifying the analogue filter.

When designing the SACD format, Sony bizarrely chose 1-bit DSD. Although it can achieve slightly higher dynamic range than CD (below 20 kHz), the improvement is slim, and creating the bitstream is difficult, as is working with it (even simple cutting/joining can cause nasty clicks). 24-bit PCM at 48 kHz would have been a much more sensible choice. As if the DSD format wasn't cumbersome enough, they then added so many layers of copy protection that had they been mattresses, the fairy-tale princess would surely not have noticed the pea of music hidden underneath. In the decades since, advances in IC technology have shifted the sweet spot for the D/A conversion stage to higher frequencies and more numerous levels, leaving DSD looking increasingly antiquated if not risible.