SACD fundamentally flawed?

"Either way, I think I'll sit on the sidelines, before spending any more money"

Bruno Putzeys also wrote:

"My position is that DVD-A is technically better by far. Either the sonic
difference will be zip, or it will be to SACD's disadvantage (specifically
when the noise drives your amplifiers nuts).

No need to sit on the sidelines if the above is true.

WVK

 

Yet more thanks to thomh for posting Mr. Putzeys' statements verbatim. Having immersed myself in AES and IEEE papers on the subject recently because of (hopefully) an upcoming DAW purchase, these postings have reinforced the opinion I've been arriving at over the past few months; that is, SACD is a wonderful delivery format but in the studio, where processing power and storage is less of an issue, PCM is a better choice.

 

I hate to say it, my friend, but I don't think one gets much benefit from SACD on the 563. DVD-A playback sounds very nice, but SACD just ain't where this player shines. At least it does get you into multichannel SACD, but even that doesn't sound all that great on this player. IMHO!

Now, why you hear smearing on many SACDs but not on Steve's is a mystery to me so far. Steve's Zombies disc is extraordinary, and I love his CCR a lot, and I have to say that the Buddy Rich and Cal Tjader discs are also terrific, but I hear great stuff on many other SACDs as well.

Just for fun, can you cite three or four SACDs where you hear a lot of this smearing? If I have one of the titles, I could give it a listen and tell you what I hear. And perhaps some other folks would also chime in, and we could see how "in synch" we are, player to player.

 

Bruno's statements are more complex and nuanced in the longer Q&A, I think, so much so that he ends up endorsing SACD as a delivery medium and DSD as a recording (but not production) medium, saying that both are a step forward sonically, but in a way that raises very difficult problems in the production chain.

So the question of which way to go is still a tough one. I'm satisfied, though, that Bruno thinks well-mastered SACDs that do not come from a DSD production chain are a step forward from CD. His conclusion is that "In terms of sound quality I believe both SACD and DVDA to be good formats for releasing remasters on. As said before, my problem with SACD lies in the practical side of affairs." So folks who have bought into SACD shouldn't think they've made a mistake, necessarily. Ditto with DVD-A. (I suspect the real choice comes down to titles and availability, and DVD-A seems to have gone to sleep here. I hope they awaken!)

 

Ok, let me get back to you in the new year as I'm kind of busy until Jan. 3rd. Too much food, visiting, etc.

 

For a good example of top-end smearing on SACD, check out The Police - Outlandos D'Amour. It's particularly obvious on "So Lonely" and "Next to You". In fact, pretty much any time a high hat or ride cymbal is played you'll hear the smear. It's quite unnatural. I think it's really difficult to listen to this album on SACD.

 

Here's my take as an engineer who has done work in DSD and DVDA, mostly at 24/88.2 and 24/96.

Comparisons are tough between the two because implementation quality varies significantly among A to Ds and playback devices and as always the recording chain simplicity and mastering quality are big factors.

Some observations:

1. While I like both done right, DSD seems just better to me. I have heard the DSD of a Meitner box and been able to compare to rela studio events and DSD is simply very analog and realistic when done right.

2. DVDA is a huge change over redbook as well, but I find that cymbals seem to sound bright and digital unless the sampling rate is 24/192. I think that faster sampling rates definitely improve transient note capture.

3. Complaining about some editing stations interim use of 352khz PCM is sort of silly as at that sampling rate you are really getting a fast sampling rate and few systems could probably highlight the difference. Plus, this is going away as better editing stations (several with no PCM conversion) become available.

4. All else being equal, I find recently done "pure DSD" recordings really highlight the beauty of Super Audio. Alison Krauss and Union Station Live and the Red Rose Sampler are stunningly real when played back on a good system.

5. Transfering analog tapes to DSD also sounds good. When Barry Wolifson went from tape to DSD (all Chesky Records I am told by Barry), we got a great sound in my opinion. I feel somewhat confident in saying this since I worked on New York Reunion and know that the SACD got me closer than the redbook in spades.

6. The controversy around the upper band noise in SACD is interesting to say the least. If you talk to John Atkinson as I have, he says he can't hear it. I can't hear it either and neither can Levinson. I wonder if this started due to some early tests that were misleading like the Warner consultant's test that was presented earlier in the SMR link as being a "unbiased source". The top DVDA investor is far from unbiased and Craig works for Warner.

7. Neither format is perfect. The real question is what software is available and how accessible is it to record. SACD vastly outnumber DVDA titles, but I have both just so I can listen nicely to Steely Dan and Donald Fagen in hirez.

8. It's really early in the game. Steve and Kevin are doing amazing things for APO and Audio Fidelity, but sonics will get even better as better chips and other implementations get created. Anyone who remembers early redbook knows where I am coming from.

I think audio life is really good these days. Lots of hirez, in both digital and vinyl forms!

 

Maybe you have to be in your early twenties.

Ah, to be young again...or at least a canine. But my dog only likes the Red Hot Chili Peppers (I'm assuming his howls are in approval).

(no, he really does howl to RHCP, exclusively)

 

Steve,

The reason I mentioned Outlandos D'Amour was that it sounds better to me on the vinyl than it does on the SACD. Perhaps the SACD has more resolution than my vinyl pressing, and makes the limiting artifacts more noticeable to me.

Technically speaking, I'm a SACD dunce, and I refrain from participating in format wars. I'm just commenting on what I am hearing, and it's easier for my ears to handle the top end on the vinyl than on the SACD. FWIW, I don't hear the smear on any other SACD release that I own.

It would be interesting to hear people report more examples of this phenomenon. I'd like to know if this is real or imagined. I have a Sony DVP NS-500V, a good budget SACD player. It could be my player, although other SACD titles do not turn me off like the Police release.

Thanks, Robert

 

Was able to squeeze in some listening time...

Roxy Music: Avalon (killer overly bright with some sort of limiting being used. Just eliminating the top end sheen and extensions of cymbal crashes. No breath of life, just a continuous wall of same level sounds!). IIRC I'm almost certain that the original didn't sound this bad. Must get a W. German pressing for comparison.

Rolling Stones Sampler: Track by track 1. Bob Dylan: Simple Twist Of Fate The Harmonica smears into distortion and waaay too bright (just this version?) 2. Pink Floyd: Money. Cymbal crashes sound fake, dull, and smeared. Snare drum sounds squashed, but the vocals sound right. The Sax solo actually hurts, same with the guitar solo! Way too bright! 3. The Who: Pinball Wizard. Now this is the worst so far. Talk about tinny sounding. 4. Herbie Hancock: Watermelon Man. That Flute is killing me arg! Just completely unpleasant IMO. Again, no breath of life. WOS. 5. Nora Jones: Come Away With Me. The "human voice" sounds normalized/synthetic and completely digitized. The cymbal brushes sound smeared and artificial to me. 6. Elton John: GBYBR. The top end of Elton's voice is not well defined like on other issues of this tune. Vocal nuiances are very smeared. Wow, did they ever trash this song, the whole thing bites IMO. The ear pain. 7. Billy Joel: Movin' Out. This is one of the best on this Sampler. The cymbal crashes sound smeared when compared to original CD pressings that are very well defined IMO and just sound more like the real thing. 8. Miles Davis: Blue In Green. Sounds very convincing. Very real sounding. Does Steve do this one? 9. Aerosmith: Sweet Emotion. The drums sound buried (smeared) and the whole song sounds like it has been maximized when compared with the Mastersound gold CD.

Does that help Gardo, sort of along the lines you were looking?

 

LeeS, Bruno (BP) comments as follows:


LeeS:
While I like both done right, DSD seems just better to me. I have heard the DSD of a Meitner box and been able to compare to rela studio events and DSD is simply very analog and realistic when done right.

BP:
Admittedly there are currently no PCM chains that do exactly as I'd like to see it.
Apart from that I strongly suspect that if Meitner had joined the DVDA camp and done 192/24 converters instead, that would have been the clear winner. The sonic impact of good electronics still is greatly underestimated.


LeeS:
DVDA is a huge change over redbook as well, but I find that cymbals seem to sound bright and digital unless the sampling rate is 24/192. I think that faster sampling rates definitely improve transient note capture.

BP:
Whether 96/24 is enough to obtain full sonic transparency is still open to some debate. However, before deciding this, we should at least do it right.


LeeS:
Complaining about some editing stations interim use of 352khz PCM is sort of silly as at that sampling rate you are really getting a fast sampling rate and few systems could probably highlight the difference. Plus, this is going away as better editing stations (several with no PCM conversion) become available.

BP:
"No PCM conversion" only allows editing. Any processing inherently entails conversion to PCM. DSD should be viewed as a PCM signal with wordlength 1 and sampling rate 2.8224MHz. As such, it behaves like any other PCM signal: processing by necessity adds word length. A straight addition of two signals of n bits yields a signal of n+1 bits. A multiplication of two signals of length n and m (e.g. a gain factor and an audio signal) yields a signal of length n+m. In PCM, the result is brought back to the nominal word length by dithering (noise shaped dither is avoided except for the mastering dither when the word length is short, e.g. 16 bits). The extremely low noise levels of 24 or 32 bit PCM make you'll have to do thousands of processing steps before the noise becomes noticeable. In DSD, dithering is no option of course. You need the aggressive noise shaping in order to obtain acceptably low noise in the audio band. The very high no! ise levels outside the band compromise the operation of subsequent noise shapers. After 5 rounds, a silent signal is enough to overload it.
So after doing anything your 1-bit signal has become much longer than 1 bit. You'll have to get it back to 1-bit using the same noise shaping that's used in the converter. Time and again.


LeeS:
All else being equal, I find recently done "pure DSD" recordings really highlight the beauty of Super Audio. Alison Krauss and Union Station Live and the Red Rose Sampler are stunningly real when played back on a good system.

BP:
Pure DSD is fine - if you don't do anything to the signal except edit it and press a disc. Heck I'm designing a converter box for friends to do just that.


LeeS:
Transfering analog tapes to DSD also sounds good. When Barry Wolifson went from tape to DSD (all Chesky Records I am told by Barry), we got a great sound in my opinion. I feel somewhat confident in saying this since I worked on New York Reunion and know that the SACD got me closer than the redbook in spades.

BP:
I don't think anyone was arguing that 44.1/16 was enough.


LeeS:
The controversy around the upper band noise in SACD is interesting to say the least. If you talk to John Atkinson as I have, he says he can't hear it. I can't hear it either and neither can Levinson. I wonder if this started due to some early tests that were misleading like the Warner consultant's test that was presented earlier in the SMR link as being a "unbiased source". The top DVDA investor is far from unbiased and Craig works for Warner.

BP:
I don't hear the outband noise either. It's a practical burden, not a (psycho)acoustical one.


LeeS:
Neither format is perfect. The real question is what software is available and how accessible is it to record. SACD vastly outnumber DVDA titles, but I have both just so I can listen nicely to Steely Dan and Donald Fagen in hirez.

BP:
Hence my comment that "SACD is probably here to stay".


___________
Thom

 

Thanks, Dave: that's exactly the kind of thing I was looking for. Here's a quick response, with more to come after cartoons are over and I get access to the main stereo again:

I've got three pressings of Avalon: the first (W. German) issue, the remaster from a few years ago, and the SACD. The first issue is mastered at a very low level--haven't listened to it for a while, but I remember it sounded kind of wimpy. The redbook remaster is much hotter but it's also compressed and a little fatiguing to my ears. The SACD stereo layer has the punch of the redbook remaster (probably some analog limiting there too) but it's much, much smoother and more detailed. Not fatiguing to listen to, IME. And I don't hear anything like the smearing you're describing.

I don't have the RS sampler so I can't be confident that the 2-channel SACD tracks on there are the same versions I have on my standalone SACDs, but just in case ... 1. Have it but haven't listened yet, 2. the 2-channel layer of DSOTM never has sounded good to me, not because of DSD but because they've eq'd and squashed it--mch is the only way to go with this release IMO, 3. "Pinball Wizard" is anything but tinny on the Tommy SACD 2-channel, and it's positively bass-heavy on the mch mix, 4-5. Don't have 'em, alas, 6. my SACD 2-ch mix of GYBR sounds to me like a hi-rez version of the original MCA release: punchy, toppy but not too bright, not as much depth as Steve's mastering yields, but still very listenable (and very nice in mch), 7. I don't hear any smearing at all on my Stranger SACD--the production quality varies on that album, probably by design to give each song a certain "sound," but the highs sound fine to me, 8. Don't have it (yet), and 9. this track is probably the best-sounding on a fine-sounding SACD of Toys in the Attic--I don't know the 2-ch mix as well as the mch, but I don't remember the drums being buried or the sound being smeared at all.

FWIW, I strongly suspect you're hearing limitations imposed by the 563a. The next question, though, is why some SACD tracks do sound good with your setup.

One SACD where I don't hear anything like the depth and transparency I'd expect is the Stan Getz Pure Getz SACD from Concord Jazz. I don't hear smearing, but it does sound as if a thin gauze sheet has been draped over the sound--a little dulled, distant, and muffled.

 

Thanks for keeping the conversation going with Bruno, Thom. This is really great, thought-provoking stuff. I've been especially interested in the transparency vs. neutrality idea that Bruno brought up with regard to vinyl. I'd not considered the differences in exactly that way, but it certainly tallies with what Steve's said about the resolution of vinyl, especially when one's listening to the direction, quality, and decay of the reverb triggered by the snare hits at the beginning of "Fortunate Son."

Bruno's idea gives me a better way to think about what I like/dislike about vinyl and CD. Cool.

 

Is the 563a one of those decks that convert DSD to PCM?

 

The Stranger SACD stereo sounds good to me - I compared it to my early Columbia Record Club CD, the remaster (ugh), and the tracks off my early Greatest Hits CD. Haven't heard an early Japan pressing though. The Stranger SACD made me go ahead and get the 52nd Street SACD, which I found very disappointing.

I was impressed by how good the stereo Miles Davis sounded - my Mastersound gold CD wasn't even close. That track is from Kind Of Blue, which I don't think Steve has done. Based on this track, I ordered the SACD...it should be here any day now.

Didn't like the stereo Elton John or The Who (prefer my UD1 for both), the stereo Aerosmith (prefer my Mastersound CD), or the stereo Pink Floyd (prefer my UD1, but I didn't compare with my "black label" Japan Harvest CD, which may be even better for this track).

 

Rumours have circled around this like crazy, but the consensus is that this player does not convert DSD->PCM.

Dave,

Sorry to hear of your unfavorable experiences with the RS sampler. I have an idea to the problem. First, let me say I think the Pio 563A is a killer player. Good video performance for the price, the SACD/DVD-A ability is almost a "bonus" IMO. If I had a place to put another DVD player, the 563A would be at the top of my list to pick up.

That being said, I think the 563A is the crutch that is preventing you from experiencing SACD the way it was supposed to be heard. You can only ask a $179 universal player to do so much. Cheap DACs, low quality analog output stages and filters don't usually translate into heavenly sound.

You may want to see if you can borrow or test a dedicated SACD deck. I enjoyed my original Sony SCD-C555ES for a long time and thought it was adequate until I got my SCD-XA777ES. It was like getting a whole new system! I'm not saying you need to jump all the way to the $2K US players to make you appreciate SACD, but I promise you will hear a difference to your liking as you go up the SACD hardware ladder.

 

Thanks guys. If it were as simple as you say, why do Steve's SACD's sound great when compared to these in your face mastered SACD's?

Sam, you are no doubt in my mind right that I could perhaps appreciate the complete SACD realm better with a dedicated player. Oh well, perhaps next year the "budget" will be in a little better shape.

Gardner, we have this fantastic invention called a volume control for bringing those low-level recordings to life.

Dob, every version of DSOTM on CD or the new SACD is from the same exact "adulterated master" with one exception only. The very first Japanese pressing from Toshiba/EMI cat # CP35-3017 is a completely flat, non-equed, no compression, no digital artifacts, pressing. I never realized how futzed with DSOTM was on every other pressing until I got this one.

 

I get what you're saying. Steve's work (on any format) sounds better to me even on a boombox. There's still some good SACDs that sometimes need a nice player to take full advantage of all those bits on the disc. Some of the artifacts or tonal qualities you describe can be rectified by a different unit IMO. Ask Gary if you can drop by to give his XA777ES a listen.

 

When Mr. Putzeys states that " 'No PCM conversion' only allows editing," I think he may be oversimplifying a bit (sorry). When editing DSD, even simple butt-cuts, require a crossfade between the two bitstreams; once again, reference the AES paper "Editing and switching in 1-bit audio streams" by Reefman and Nuijten.

On the larger matter, I don't think we've really begun to experience the problems that Mr. Putzeys (or for that matter Lipshitz, Vanderkooy, and others) is warning us about yet. Much of the SACD material out there qualifies as an application of the technology which he endorses; DSD as delivery medium only- reissues produced in other formats and converted at a late stage for sale and the like. Since full-scale (24+ tracks) DSD multitrack decks have been on the market for a relatively short time and non-DAW, DSD mixers are not yet available there is little, if any, DSD material out there tracked, mixed, edited, and mastered in the (ostensibly) DSD domain. So, I don't think we've heard the worst-case scenario yet. Neve intends to release a nominally-DSD core for one of its consoles soon, though; so, with a complete production chain in place, that opportunity may not be far off.

 

Dave, do you have an original copy, or a CDR copy? I ask because I'm pretty certain that the original was pressed with the de-emphasis bit set. If you have a CDR copy, there's a pretty good chance that it has the pre-emphasis bit set incorrectly.

Also, my guess is that that particular CD was pressed from a second gen tape. Not that it'd make a huge difference - it sounds pretty nice anyways. (disclaimer: I have not yet compared the old Toshiba with the new SACD.

jeff

p.s I have an original Toshiba copy, not a CDR, BTW

 

Maybe someday I'll find one for a reasonable price.

 

Jeff, I have the "real deal" as well. (Que pasa pre-emphasis bit set?)

 

Keep looking brother Dob. I found mine on Ebay that for some reason or another was basically ignored by bidders.

 

If you try to burn a CD-R and your original CD has pre-emphasis (a lot of old Columbia CDs have them, for example - and so does the Toshiba Abbey Road CD) and you don't set the de-emphasis bit, your CD-R copy will be bright as heck...I'm talking like +10db at 10k. The idea is (was, actually, it's not used anymore) kind of like the RIAA equalization on LPs. It has to be undone by your hifi.

You can set the bit in EAC pretty easily. If you do a search, you should/may find the procedure in the archives. I've done it myself, so I can walk you through it if necessary.

I don't know how to determine if the original CD has pre-emphasis or not...other than by listening to the burned CD-R.