What's new

Help, I'm ready to buy into either SACD or DVD-A (1 Viewer)

Ian Montgomerie

Stunt Coordinator
Joined
Feb 2, 2002
Messages
112
A rather problematic assumption many SACD supporters seem to make is that the theory of DSD has much to do with what actually comes out the DACs of an SACD player. The fact is, any theoretical advantages of DSD over PCM are lost when, some time between the ADC in the studio and the DAC on your player, are lost when the music spends some time as PCM. When this happens, you really get the worst of both worlds. Because the music spent some of its life as DSD, you get flaws of DSD that PCM does not have, like inferior SNR at higher frequencies. And because the music spent some of its life as PCM, you get any disadvantages PCM might have relative to DSD.

And this basically always happens. If any post-processing on the music was done in the studio, they basically had to convert it to PCM to do it. They can only avoid that PCM stage if they do no post-processing whatsoever, it is basically straight from mike to tape (which is completely unworkable for many styles of music, and for 5.1 mixes). But even worse than that, the fact that the audio goes into the SACD decoder as DSD, and comes out as DSD, doesn't mean it was always DSD inside the decoder. Guess what, SACD decoders in players are no more able than SACD boxes in the studio to do post-processing directly on the DSD. What you have is a DSD->PCM stage, bass management is run on the PCM, then there is a PCM->DSD stage for output. (I don't know if there may be SACD decoders available which do no post-processing whatsoever, but if you have a chip which is even capable of bass management, dollars to donuts the bass management runs on PCM).

SACD as it is now is pure hype, because the theory of DSD is trumpeted over and over, but in actual recordings and actual players, any potential advantages are eliminated by PCM processing stages (while leaving the disadvantages). And it's not like this is some temporary setback. DSD domain post-processing is a whole new technological ball game. Even if SACD really takes off and people start to really invest in DSD post-processing technology, we're talking 5+ years before you'd have it be the norm in both studios and players. Without that, you're basically really limited in what you can do with the audio. Two channel only, since you won't really be able to create a good multichannel mix. No PP means most genres of popular music (as opposed to classical/jazz/etc) basically can't be recorded well, you would get no better than live-recording quality. And the player can't have bass management, so heaven help you if you have something other than two large main speakers positioned ideally, with analog-based volume control (and analog crossover for your subwoofer if you use one).
 

KeithH

Senior HTF Member
Joined
Mar 28, 2000
Messages
9,413
Ian said:

SACD as it is now is pure hype
Now that I wholeheartedly disagree with. My ears tell me differently. I don't have people attempting subliminal seduction with SACD as I listen. It's just me, and sometimes my fiancee, who is no audiophile. We both have come to the obvious conclusion that SACD is an awesome format. Is it better than DVD-Audio? That I cannot answer.
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
A rather problematic assumption many SACD supporters seem to make is that the theory of DSD has much to do with what actually comes out the DACs of an SACD player.
Technically, the DAC on the SACD player can ONLY output a DSD decoded signal translation. Again, if it was a pure DSD recording, you get pure DSD sound.
Ian, I think you need to listen the the SFO Mahler 6 or Red Rose Music sampler. They sound awesome and you will hear the vast difference over redbook CD. :)
 

JaleelK

Second Unit
Joined
Feb 28, 2001
Messages
296
I don't mean by any means that SACD is better or worse than DVD-A, I don't have the experience to back any comparison, I just didn't want to everybody in this forum to think academics are neutral people, which is very rarely the case.
As I and others who possess an objectivist view to audio have often said, the way to eliminate bias in any kind of listening comparison is to conduct of a level matched double blind listening, that is the only way.
 

Ian Montgomerie

Stunt Coordinator
Joined
Feb 2, 2002
Messages
112
This is simply not true. There are many pure DSD recordings out there which clearly highlight the benefits of Super Audio.

The problem with PCM conversions has largely been solved by new mastering workstations put out by Sony and EMM Labs.
"in actual recordings AND ACTUAL PLAYERS". I would love if you would find for me a DSD DSP which does not internally do DSD->PCM then PCM->DSD for effects processing. The companies making the decoders don't advertise that they do this, but it's there, and is one of the things you can find out from the decoder chip marketing (rather than consumer marketing). It doesn't matter what's on the disc. The decoder in the player has a processing stage involving conversion to PCM. To my knowledge there is no DSP capable of doing bass redirection, gain control, and time alignment directly on DSD data. I suppose that if you had a player with no SACD bass management hardware, you might be hearing DSD which hasn't spent time as PCM. I've not personally heard of such a beast, but then I don't have an encyclopedic database of all SACD decoders ever made.
 

KeithH

Senior HTF Member
Joined
Mar 28, 2000
Messages
9,413
Ian, where do you get this from? It is well known that the Pioneer DV-AX10 converts DSD to PCM, but where do you get the idea that this is common practice? The majority of SACD players out there have DSD DACs, period. There are a lot of people out there who have dug under the hood of SACD players and who know a lot about the parts used, and none of these people have made this accusation.
 

John Kotches

Senior HTF Member
Joined
Mar 14, 2000
Messages
2,635
Keith,

When Ian says "effects processing" I believe he is mentioning this in a recording studio context.

Philips' upcoming multi-channel SACD players will do an on the fly DSD-> PCM conversion for Bass Management and Time Alignment, then a PCM -> DSD conversion followed by a DSD D/A conversion. Philips coinvented the format and they can't do a common operation on the DSD data natively.

The original plan for Denons DV-9000 player for SACD was to transfer DSD native to the 5803 receiver, do an on the fly conversion to PCM, then apply bass management and time alignment followed by a D/A conversion. The DV-9000 was never planned to handle DSD conversion in the player.

These are interesting things, methinks.

Regards,
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
DSD for effects processing.
Again, not true. The Sony multi-channel players like the XA777 do not use any DSD-PCM conversion to retrieve multi-channel sound. The DSD process is totally complete in these DSD players.
Me thinks you need to read more at Sony's web-site about how this technology works.
:laugh:
 

John Kotches

Senior HTF Member
Joined
Mar 14, 2000
Messages
2,635
Lee,

Methinks you need to understand what Ian is talking about.....

The question at hand is either this:

Do Sony players perform a DSD-> PCM step when applying Time Alignment and Bass Management, then perform a PCM -> DSD conversion for D/A output as analog. I haven't gotten an answer from Sony, although I have asked.

Or this:
Do studio solutions perform DSD -> PCM -> DSD steps to handle common tasks like mixing from multi-track mono to stereo and/or 5.1, equalization, reverberation etc.

THere are a vanishingly small number of recordings that involve no mixing.

We don't know whether the newest Sony mixing stations are using an intermediate steps that involve PCM, as it's a black box.

Regards,
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
A rather problematic assumption many SACD supporters seem to make is that the theory of DSD has much to do with what actually comes out the DACs of an SACD player.
The above is from Ian's original post. There is simply no PCM in the conversion from DSD encoded signal in the player.
Due to the newness of DSD, the real issue is bass management and other multi-channel-related effects in the receiver, etc. because the DSP chips hence far have been geared understandably to PCM.
In other words, there are two different issues here:
1. DSD DAC decoding - the answer here is no PCM involved by definition.
2. DSP effects - being developed for DSD is my understanding but I am not 100% sure. There have been some temporary 8 bit stage processing/conversion early on but this is designed into the DSD spec and is not strictly PCM as such. For more details, see www.superaudio-cd.com
In any event for either one, the answer is evident in one's own ears - great sound! :)
 

John Kotches

Senior HTF Member
Joined
Mar 14, 2000
Messages
2,635
Lee,

Ian points to the entire signal chain as not maintaining "pure DSD" as Sony has marketed it.

In other words, not maintaining DSD from A/D input at the front end to D/A output at the back end.

Sony's solution is Wide DSD, 8-bit @ 2.8mHz. 8-bit means PCM, period. Philips' solution is 32bit @ 352kHz, also PCM.

New solutions from Sony, well that's a good question. It's pretty much a black box, and we don't know what happens inside the black box.

Regards,
 

KeithH

Senior HTF Member
Joined
Mar 28, 2000
Messages
9,413
All, I was considering Ian's comments from the standpoint of playback of stereo SACDs. Sony players do not convert DSD to PCM during playback of stereo material, which is what I listen to 99% of the time. The Pioneer 'AX10 converts DSD to PCM. As for what is happening in the studio, I have no idea. Apparently no one except a few people at Sony do either.
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
Sony's solution is Wide DSD, 8-bit @ 2.8mHz. 8-bit means PCM, period. Philips' solution is 32bit @ 352kHz, also PCM.
Not really John. 8-bit at this high sampling rate is NOT PCM, just wide DSD.
In any event, the way Keith and I read it clearly is the idea that the DAC in the SACD box is using PCM and this is clearly not the case.
Why don't you query Dr. Demery on this one at Audio Asylum? I'm sure he knows. I will try to reach David Kawakami as well.
:)
 

John Kotches

Senior HTF Member
Joined
Mar 14, 2000
Messages
2,635
Lee,

Regardless of how it's marketed, 8-bit depth is PCM. You can encode 256 discrete values for the delta of the waveform @ any point in time. You can also manipulate the data with additions/subtractions etc etc etc, which is the real reason that Wide-DSD is in use.

Neither Dr. Demery or Mr. Kawakami are an unbiased source of information... Dr. Demery was with Philips, and is now with Sony, and Mr. Kawakami is also with sony.

Regards,
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
Neither Dr. Demery or Mr. Kawakami are an unbiased source of information... Dr. Demery was with Philips, and is now with Sony, and Mr. Kawakami is also with sony.
Why do we need an unbiased source for a technical discussion? Demery and Kawakami played a major role in the R&D of DSD, so aren't they the best to state how it works?
 

John Kotches

Senior HTF Member
Joined
Mar 14, 2000
Messages
2,635
Lee,

Earlier in the discussion when I brought up Bob Stuart's paper on the characteristics required for high quality audio, you called that a biased source of information. Now, you're claiming that info from Sony (that will be SACD slanted) is not. Both of these are technical issues.

Do you see the double standard you're invoking here? I do, and I doubt I'm the only one.

With respect to Wide-DSD, once you go multi-bit you are no longer in Delta-Sigma territory. What do you think the 8-bits is measuring? It's the AMPLITUDE of the change.

As to the greater accuracy of DSD, that is a matter of opinion. Tradeoffs are made with both formats.

Further, some people think tube amplifiers sound better, even though they have substantially higher distortion. The distortion products are euphonic, being predominately 2nd and 4th order products at 1 and 2 octave intervals which makes the result harmonically richer than it was in the recording.

Regards,
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
Further, some people think tube amplifiers sound better, even though they have substantially higher distortion. The distortion products are euphonic, being predominately 2nd and 4th order products at 1 and 2 octave intervals which makes the result harmonically richer than it was in the recording.
This is not true for all tube amps, just some. The better ones can sound good and measure well also (or not). In any event, the sound is more important than measurements because you can't measure everything and you can't hear everything you can measure.
 

John Kotches

Senior HTF Member
Joined
Mar 14, 2000
Messages
2,635
Lee,

The broader discussion was the "pure DSD" marketing hype, which is where wide-DSD (and the Philips alternative) came into play. Ian's contention is that as soon as we start talking about any signal processing at all, mixing, eq, even derivation of a .1 for a multi-channel mix, then you cease to have "pure DSD".

OTOH, when talking about PCM, the signal can be captured, processed and delivered @ 24bit/96K (or higher). There need be no intermediate formats (wide DSD or 32bit/352K) involved.

Why don't you ask Ian what his credentials are, and determine for yourself if he's in a position to know what's going on.

With players, so long as we're talking no signal processing, ie Bass Management and Time Alignment, it's already been said that it is not necessary to perform a PCM transformation. All that's done on a 5.1 stream without BM and TA is a DST decompression, and a passthru to the appropriate DAC(s).

The player is but the last piece in the puzzle, and you are certainly aware of how many steps are involved between microphone input and player output on the majority of recordings.

Regards,
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
There need be no intermediate formats (wide DSD or 32bit/352K) involved.
I think this statement shades the truth, because wide DSD is not really an intermediate format as I look at it.

Can you answer this question John:

Why can't you measure the amplitude of the waveform with changes in 8-bit, or "wide DSD" as Sony would call it and also have 8-bit PCM conversion which uses height?

Are they not two different things? In wide DSD, it is the change of the wave, not the absolute value, therefore is Delta-Sigma Modulated...
 

Lee Scoggins

Senior HTF Member
Joined
Aug 30, 2001
Messages
6,395
Location
Atlanta, Georgia
Real Name
Lee
John Atkinson of Stereophile Magazine summed up the difference between DSD and PCM nicely for the Sony SCD-1 review...I think this helps move the discussion along:

The use of such a high sampling frequency would mean the ADC's analog antialiasing filter needn't be a brickwall type but could instead be a sonically benign low-order type; linearity would inherently be excellent; there would be no digital decimation filter, with its necessary mathematical approximations on either the A/D or D/A conversions reintroducing PCM quantization noise or time-domain dispersion problems; there would be no multi-bit DAC, with its possible performance compromises—this would be the closest thing to a digital topology with analoglike properties.
 

Users who are viewing this thread

Top