2xDSD - Smoke and mirrors, or will there appear better masterings on it?

[Nightlord]

As the question of this was too tenderfooted for a sideline in the CES report, here's a thread.

Most interesting trend:

The increasing number of DACs and CD players that will support 2X DSD.

Has anyone seen any statements from the music industry that they are going to supply less dynamically compressed mastering on this format, or is it just another smoke and mirrors trick
for all but a very few special pressings? Current trend isn't positive with new releases of Beatles/Beach Boys etc that's less dynamic than previous releases, even if they are claiming to be gentle with the source.

Can we expect better transfers from old mastertapes and better new masters than with the current 10-12dB crest dynamics, that EASILY fit the CD format and then some...

 

[opus111]

Isn't dynamic compression built into the format? It was back in 2000 I believe when Stanley Lipshitz pointed out in his paper that 1bit systems can never be properly dithered, noise modulation is always the result of an improperly dithered quantizer.

 

[LL21]

BTW, i know the CD 'limit' is 90db or so...but what in practice is considered the typical maximium dynamic range of classical music (which i presume tends to have the largest dynamic range most of the time)?

 

[opus111]

I tend to think less than 60dB as the low frequency hall ambience is quite noticeable when a recording starts, above this level.

 

[Nightlord]

Isn't dynamic compression built into the format? It was back in 2000 I believe when Stanley Lipshitz pointed out in his paper that 1bit systems can never be properly dithered, noise modulation is always the result of an improperly dithered quantizer.

You're quite far off target. While you could have a point, it's on a quite different (much smaller) scale than what occurs due to "music business requirements" compression levels.

 

[Nightlord]

BTW, i know the CD 'limit' is 90db or so...but what in practice is considered the typical maximium dynamic range of classical music (which i presume tends to have the largest dynamic range most of the time)?

So... we can do 90dB.... and the music the industry puts in in good cases uses a (average +/- 12dB ) 24dB range. The largest crest I've measured on any CD I have is 27dB... yielding 54-55dB usage of the CDs 90...

So I can't see why we really needs a better format, if we don't get better 'data' to put into it. The chance with a new format is that we might get it, but will we? Or will it still be the same unlistenable compressed crap with a higher resolution in time - which really doesn't do anything to make it become listenable, probably only highlights the problem further if anything.

 

[Mike Lavigne]

As the question of this was too tenderfooted for a sideline in the CES report, here's a thread.



Has anyone seen any statements from the music industry that they are going to supply less dynamically compressed mastering on this format, or is it just another smoke and mirrors trick
for all but a very few special pressings? Current trend isn't positive with new releases of Beatles/Beach Boys etc that's less dynamic than previous releases, even if they are claiming to be gentle with the source.

Can we expect better transfers from old mastertapes and better new masters than with the current 10-12dB crest dynamics, that EASILY fit the CD format and then some...

i have approx 150 2xdsd files on my server. almost 100% of them are RTR master tape sourced with zero compression. i see the most viable source going forward for 2xdsd to be analog tapes. as far as the current recording industry adapting direct to 2xdsd as a normal recording format i have little expectations. but really, i don't care much about it since the music is not that attractive to me mostly.

and i don't see much 'market pressure' to push studios and artists to improve the sound, other than niche audiophile labels.

 

[Nightlord]

i have approx 150 2xdsd files on my server. almost 100% of them are RTR master tape sourced with zero compression. i see the most viable source going forward for 2xdsd to be analog tapes. as far as the current recording industry adapting direct to 2xdsd as a normal recording format i have little expectations. but really, i don't care much about it since the music is not that attractive to me mostly.

and i don't see much 'market pressure' to push studios and artists to improve the sound, other than niche audiophile labels.

How do you get those, unless you're in the business? Doesn't sound like anything I'll be able to find on town...

We have to all help out, that's part of why I made the original comment and then this thread... to push awareness and help put pressure on the studios...

 

[Mike Lavigne]

How do you get those, unless you're in the business? Doesn't sound like anything I'll be able to find on town...

We have to all help out, that's part of why I made the original comment and then this thread... to push awareness and help put pressure on the studios...

it's not who you are, it's who you know.

i've been around when a few of them have been recorded, a few were done off my turntable. and like many of us, i know people who know people. i also own a good number of quality 15ips 1/4" and 1/2" master tape dubs which could be used as a source although i've not personally made any 2xdsd files from those.

anyone with a $2000 KORG can potentially do that. not saying that all 2xdsd files are created equal....because they are not. some hardware manufacturers have a few of these files they've shared with their customers for demo purposes.

i've not yet seen any commercially avalable 2xdsd yet; but we are on the ground floor of this format so hopefully that is the next thing.

the best thing you can do to push this along is to ask for and purchase dsd downloads from the various sources. once that gets rolling then 2xdsd is next....since lots of the dsd is sourced from 2xdsd files made from tapes.

get a 2xdsd capable dac and jump into dsd downloads, and forget about the recording industry catching onto this anytime soon. if that ever happens great, but don't hold your breath.

 

[Nightlord]

forget about the recording industry catching onto this anytime soon. if that ever happens great, but don't hold your breath.

ok. Close to useless as a selling point for gear then. Thanks.

( If you can get hold of original masters or pre-mastering tapes, then you can just at well make better CD-transfers, but that isn't happening... )

 

[Imperial]

So... we can do 90dB.... and the music the industry puts in in good cases uses a (average +/- 12dB ) 24dB range. The largest crest I've measured on any CD I have is 27dB... yielding 54-55dB usage of the CDs 90...

So I can't see why we really needs a better format, if we don't get better 'data' to put into it. The chance with a new format is that we might get it, but will we? Or will it still be the same unlistenable compressed crap with a higher resolution in time - which really doesn't do anything to make it become listenable, probably only highlights the problem further if anything.

The ear has a dynamic range of some 130db to 140db, depending on where you read about it. But, that's not to say that a recording with that range will have any benefit at all.
Maximum detectable speech/song variation is some 40db dynamic range, and music made by instruments in a venue concert hall rarely exceed 80db dynamic range. (wikipedia sort of science..) so it seems it all comes down to noise floor, and how you digitize.
In the listening enviroment you have that noise floor in the room, and again in the recording itself, as noise overlaid into the music.

So when we listen we hear two noise floors interacting, and then there is the noise floor of the playback chain, making it 3 actual noise floor components, again, interacting.
We can listen deeply into a noise floor, so analogue, because it does not really brick wall the noise, contain much more information that digital wastes, filters away, but the was lot of musical information there!! Now it is gone!

When they some day find a way of doing the noise floor the way it is on say a vinyl or on a Tape, then digital will sound much better!
You only need say in excess of 80db dynamic range, the real challenge, is what do you do about the noise floor, you should not remove it, because the ear can perceive all the way down to 0db... and all the way up to 140db...

We do NOT need a better format, we just don't! I totally agree.
What ever they do, it will most likely involve a better way of handling the noise that will always permeate music.
Our brains handle and sort out this noise, still letting us hear the music.
I'm not so sure that digital makes that easier, I'm thinking that something happens when you digitize, when you sample, that the brain is able to detect, and that screws up that built in separator, music from noise, that the brain has!

I'm quite sure that someone will be able to crack the code, how to fool the brain, dynamic range is just fine and dandy, but when you recreate digital back to analogue, it sort of isn't the same as what was put in.
the brain knows, somehow.

PCM, DXD what ever the format, the smoke and mirrors are most likely what we need, to fool that brain into thinking and then decoding, as if it still was analogue.

I'm thinking THIS is the challenge at hand (ear..) not better resolution..

Imperial.

 

[opus111]

Indeed we do not need a better format, we just need to make better use of the existing format. Besides, DSD is a worse format than RBCD.

 

[FrantzM]

Indeed we do not need a better format, we just need to make better use of the existing format. Besides, DSD is a worse format than RBCD.

I would appreciate your explanations why. Thanks in advance.

 

[Mike Lavigne]

Indeed we do not need a better format, we just need to make better use of the existing format. Besides, DSD is a worse format than RBCD.

so you've done plenty of listening tests of dsd, 2xdsd, and redbook CD compared to a source thru SOTA gear and come to that conclusion?

i'm not trying to be sarcasitic. i'm only trying to determine whether it's theory or performance which you base your opinion on. for me it's performance......on a regular basis i do these format comparisons in my own system. and i can compare them directly to the source tape, or Lp.....and easily hear my preference.

i like redbook very much. and higher sampling rate PCM too. i just like dsd/SACD better, and then 2xdsd better yet. so it's not that i'm biased toward one of them. it's just listening.

 

[opus111]

I would appreciate your explanations why.

You might begin with this paper by Stanley Lipshitz.

http://sjeng.org/ftp/SACD.pdf

The short version for those who are math-averse is this: a 1bit quantizer can't be dithered correctly, and correct dithering is essential for any digital system to be linear.

 

[opus111]

so you've done plenty of listening tests of dsd, 2xdsd, and redbook CD compared to a source thru SOTA gear and come to that conclusion?

Listened to various RBCDs which have used 1bit systems (both bitstream and DSD) in the recording process. I hear what Stanley Lipshitz talks about in his paper (just cited above) - noise modulation. Have not listened to DSD128 but see no reason why it would be different as its still just a 1bit system which Lipshitz shows is fundamentally flawed.

i'm not trying to be sarcasitic. i'm only trying to determine whether it's theory or performance which you base your opinion on.

Its not my opinion here, its the contents of my perception - listening. Theory says it can't work, practice shows it doesn't in fact work. So the two are in harmony.

for me it's performance......on a regular basis i do these format comparisons in my own system. and i can compare them directly to the source tape, or Lp.....and easily hear my preference.

Well I'm curious here - what DACs are you using to listen to RBCD on? My hypothesis (which I'm more than willing to test right here) is that your chosen DACs aren't doing justice to the RBCD material.

 

[Imperial]

Well I'm curious here - what DACs are you using to listen to RBCD on? My hypothesis (which I'm more than willing to test right here) is that your chosen DACs aren't doing justice to the RBCD material.

And how would you test that...

Now I'm curious..

Imperial.

 

[microstrip]

i have approx 150 2xdsd files on my server. almost 100% of them are RTR master tape sourced with zero compression. i see the most viable source going forward for 2xdsd to be analog tapes. as far as the current recording industry adapting direct to 2xdsd as a normal recording format i have little expectations. but really, i don't care much about it since the music is not that attractive to me mostly.

and i don't see much 'market pressure' to push studios and artists to improve the sound, other than niche audiophile labels.

Mike,

It seems you have a very valuable asset. Can you tell us something about the system that was used to digitize these RTR master tapes?
I hope that you good backups of these files, but I think that for increased safety you should keep a backup in a small country of western Europe.

As far as I know (but I may be wrong, I hope that the experts will debate this aspect) the big issue with DSD is that there are no mastering tools in DSD - the DSD recordings are converted to analog or PCM for mastering. Recording in pure DSD must be something like recording a direct cut LP.

 

[Imperial]

Right, firstly, you can master DSD directly, say with Merging technologies Pyramix gear.. and Sony, of course, has a platform for it!
So lettuce not start saying that one cannot do this!!!

Imperial.

 

[FrantzM]

Opus

This is my contention as well based on what I think I know about signal processing, not much I must say:0 Thanks and this is a discussion I would like to pursue. As for your short version any shorter than that and it would require only one bit

I would like an honest discussion on the subject so that to answer some questions, among these:

Why so much energy into developing SACD
Why is it that some people maintiain it sounds better than PCM? I am assuming identical mastering, a really overly optimistic assumption but ...