Music Player Network
Previous Thread
Next Thread
Print Thread
Page 6 of 6 1 2 3 4 5 6
Joined: Oct 2000
Posts: 4,223
C
MP Hall of Fame Member
OP Offline
MP Hall of Fame Member
C
Joined: Oct 2000
Posts: 4,223
Quote:
Eric,
From my understanding (which is admittedly a little thin at this point in the conversation) there is an inherent benefit in multibit at the same sample rate, but the one bit DSD is at a much faster rate than multibit oversampled PCM, to the point that I think that the drawbacks don't compare.
8x oversampled, but multibit vs. 64x oversampled but one bit. Is there an advantage? I don't know. I was led to believe that they could be effectively the same.
Nika.


Nika,

That trade-off only works until you get down to 2 bits, which is the number of bits you need to properly dither, and that's where you get into trouble in other areas. Yes, you can still juggle the numbers at 2-bit or 1-bit if you like, but if the correct dither (white noise with a triangular pdf and a width of 2 LSBs) covers the full range of the converter, the convertor overloads upon input. That's why, normally, when you hear of low-bit convertors, you hear of as low as 3-bit convertors, but usually not 2 or 1.

So, yes, the 1-bit convertor may have a higher sampling rate, but that's not going to help the other problems you get with a 1-bit DSM, such as the non-random cyclic HF noise, the resultant 3rd harmonic distortion, and the excessive HF pressure which, although is above the range of human hearing, will no less produce enough energy to burn the shit out of the tweeters in your monitors - all of these factors are beyond fixing with high sampling rates, and must be dealt with in other ways. And they are, no doubt, but in ways that significantly degrade the audio, and that is why the 1-bit convertor is considered obsolete today.

I hope this helps.

Eric \:\)


Eric Vincent (ASCAP)
http://www.curvedominant.com
Joined: Mar 2001
Posts: 2,938
N
MP Hall of Fame Member
Offline
MP Hall of Fame Member
N
Joined: Mar 2001
Posts: 2,938
Quote:
Originally posted by Curve Dominant:


Nika,

That trade-off only works until you get down to 2 bits, which is the number of bits you need to properly dither, and that's where you get into trouble in other areas. Yes, you can still juggle the numbers at 2-bit or 1-bit if you like, but if the correct dither (white noise with a triangular pdf and a width of 2 LSBs) covers the full range of the converter, the convertor overloads upon input. That's why, normally, when you hear of low-bit convertors, you hear of as low as 3-bit convertors, but usually not 2 or 1.

So, yes, the 1-bit convertor may have a higher sampling rate, but that's not going to help the other problems you get with a 1-bit DSM, such as the non-random cyclic HF noise, the resultant 3rd harmonic distortion, and the excessive HF pressure which, although is above the range of human hearing, will no less produce enough energy to burn the shit out of the tweeters in your monitors - all of these factors are beyond fixing with high sampling rates, and must be dealt with in other ways. And they are, no doubt, but in ways that significantly degrade the audio, and that is why the 1-bit convertor is considered obsolete today.

I hope this helps.

Eric \:\)


Eric,

This is not correct. Dither is not applied at this stage. Once the data is at this point dither has already been applied, so having enough extra bits at the tail end of the converter to dither prior to conversion to analog is unnecessary. I think you might be confusing DSD as a form of recording/editing where it falls weak specifically because of the dither problems within 1 bit signals, and DSD as a delivery format where the signal has already been dithered.

As for the other issues you mention with mostly HF problems in DSD signals, this is being dealt with with analog filters in playback equipment that are well above the hearing range such that they should have no effect on hearing, but low enough to remove the problems you're speaking of. Hutch from Manley addressed manufacturing concerns with equipment that is intended to play back HF DSD data, and I believe they design their equipment specifically to handle those issues accordingly.

I'm not here defending 1 bit converters - you don't seem in any way interested in engaging in conversation on the matter anyway. Your manner in this thread has become suspiciously hostile and agressive, mocking and attacking people who have attempted to answer your questions, and then faulting them for not taking more time to engage the conversation with you.

I'll remind you that you asked me a very simple question: "what is the supposed benefit of using DSD as a delivery format". I guess I'll just restate my answer and leave it at that. I'm not DEFENDING why it is valid. I'm merely presenting a possibility to you for the sake of your exploration. You did at least APPEAR to be asking these questions for the sake of exploring. I'm not convinced anymore that that was your motive.

The answer I was presenting to your question was "to remove the digital part of the reconstruction filter from the consumer's equipment, essentially giving them the data one stage later in the digital audio conversion process."

I'm going to leave this thread alone for now. I see no reason to be asked what appears to be a genuine question and give a very long and comprehensive answer to the nature that I believed I was being asked, only to be mocked and essentially "reprimanded" for giving such an answer. It seems you have also alienated and insulted others who were attempting to contribute as well. I hope you find the answers you're looking for at some point, if indeed you're looking for answers. I believe you're actually looking for something else.

I'll see you on the next thread,
Nika.

Joined: Oct 2000
Posts: 4,223
C
MP Hall of Fame Member
OP Offline
MP Hall of Fame Member
C
Joined: Oct 2000
Posts: 4,223
Quote:
I'm going to leave this thread alone for now. I see no reason to be asked what appears to be a genuine question and give a very long and comprehensive answer to the nature that I believed I was being asked, only to be mocked and essentially "reprimanded" for giving such an answer. It seems you have also alienated and insulted others who were attempting to contribute as well. I hope you find the answers you're looking for at some point, if indeed you're looking for answers. I believe you're actually looking for something else.
I'll see you on the next thread,
Nika.


Before you go, do you mind explaining what is it, exactly, do you believe I'm looking for? I don't work for (or sell products of) any rival to Sony or Phillips, if that's what you're implying; I'm a freelance composer, curious about emerging formats. That's it.

Trying to get empirical data out of the pro-SACD camp can be a frustrating experience, as you have noted often yourself, Nika. What we've clearly established from that exercise is: "Super Audio" it's not. If I've been looking for an answer, and that's the answer, then yes, we've reached the point of redundancy.

I'm not here defending 1 bit converters - you don't seem in any way interested in engaging in conversation on the matter anyway. Your manner in this thread has become suspiciously hostile and agressive, mocking and attacking people who have attempted to answer your questions, and then faulting them for not taking more time to engage the conversation with you.

That is incorrect. What I faulted was a lack of specifics, amidst an over-abundance of rhetoric, but you know that, because you complained about that yourself earlier in this thread. Do I have to get crafty1 back here as a reminder?

As for the other issues you mention with mostly HF problems in DSD signals, this is being dealt with with analog filters in playback equipment that are well above the hearing range such that they should have no effect on hearing, but low enough to remove the problems you're speaking of. Hutch from Manley addressed manufacturing concerns with equipment that is intended to play back HF DSD data, and I believe they design their equipment specifically to handle those issues accordingly.

I don't remember seeing Hutch posting anything to that effect. Manley makes front-end processors, not consumer gear. And I don't see how analog filters will make it a competing product to PCM products. If the high frequency problems inherent in 1-bit signal are not inherent in well-designed PCM, why switch?

I'm not trying to "mock" you, Nika, I'm simply trying to understand how "super" the "Super Audio CD" really is. If you're saying that, theoretically, it could possibly work just as well as today's CD, fine! I'll take your word for it, and I apologize if I offended you, but maybe you shouldn't have taken this so personally.

On that note, if I suddenly start seeing SACD players in the Sweetwater catalogs, I'm going to get awfully suspicious! \:D

Eric \:\)


Eric Vincent (ASCAP)
http://www.curvedominant.com
Joined: Oct 2000
Posts: 4,223
C
MP Hall of Fame Member
OP Offline
MP Hall of Fame Member
C
Joined: Oct 2000
Posts: 4,223
A post-script to my previous post:

Nika called me here at the Curve Lab to remind me that, yes, there are some lucky souls who count Manley gear as components in their home audio systems. I have to point out, however, that if your home-system is not an all-Manley system (which it can't be because Manley doesn/t manufacture stereo speakers), any non-linearity in your system can reproduce aliasing error created by 1-bit conversion, even if those artifacts originally occurred outside of the passband.

Nika also asked me to tone it down a bit, which I will. \:\) Things appear harsher in print than they are originally intended; I have to improve my dithering ratio in that regard.

E \:\)


Eric Vincent (ASCAP)
http://www.curvedominant.com
Joined: Feb 2001
Posts: 87
Senior Member
Offline
Senior Member
Joined: Feb 2001
Posts: 87
George asked me to clarify something here:

Tom Jung was playing back some SACD material with the player's internal LP filters disengaged and ran into some "chirping" sounds coming through his Manley 50w monoblocks from all this unfiltered HF energy. This was because these amplifiers are basically flat to 200KHz and will play back the spurious high frequency noise that rings down. So we made some different external passive LP filters for Tom to play with and also had him try a 400pf cap in the amplifier's feedback circuit to achieve the high frequency roll off required in two different ways. That all worked. And worked well.

I find it goofy that we, as a playback gear manufacturer, have worked hard over the years to get better HF response out of our tube gear, and specifically our output transformers, but now we are being asked to roll all that achievement off because of some inherent problem with SACD and the players' manufacturers inability to make nice sounding filters...

So, ok. If someone needs one of our amplifiers to be 20dB down at 50KHz, we can comply with the request... or if someone needs some pretty inaudible external LP passive filters made, we have done that too.


EveAnna Manley, Manley Labs
manleylabs.com
Joined: Oct 2000
Posts: 4,223
C
MP Hall of Fame Member
OP Offline
MP Hall of Fame Member
C
Joined: Oct 2000
Posts: 4,223
Thank you EveAnna for that cool bit, although I have to apologize for chuckling at this part (because I know it's not really funny, it kinda sucks actually):
Quote:
I find it goofy that we, as a playback gear manufacturer, have worked hard over the years to get better HF response out of our tube gear, and specifically our output transformers, but now we are being asked to roll all that achievement off because of some inherent problem with SACD and the players' manufacturers inability to make nice sounding filters...


Well, I just returned from my very first reliable experience A/B'ing SACD with Redbook CD, at a big name hi-fi store here in downtown Philadelphia which I will refrain from naming, in order to spare the store manager further humiliation...LOL!

I started out playing dumb: "So, why's SACD different?" First off, he couldn't properly pronounce "pulse code modulation," and mangled it a completely different way each time he attempted the phrase - that was funny enough. Then he told me that DSD is "pure digital: all ones and zeroes, PCM is not." So I asked him, "Pcm's not ones and zeroes?" He shook his head no, and said, "It's agalithms." I said, "You mean algorithms?"...this is the store manager, mind you, and it went on like this for some time, and my face was in excruciating pain trying to suppress an imminent explosion of laughter.

Finally a younger dude, who at least knew what a sampling rate was, walked over to save his boss from a complete massacre. Young dude had Miles "Kind Of Blue" on a SACD and Redbook (on one disc, that is), so into the isolated and treated room we went.

The player was a Sony (duh), I didn't catch the model number, but it was going for $1700. Speakers were Klipsch, $1500. Young dude puts on the Redbook first..."So What" opens (oh, what a gorgeous intro to a song), and we let it play through Miles' first solo. Then, same disc, same system, no settings changed, young dude made the switcheroo to SACD.

I could tell the difference, almost immediately, before Miles' solo.

The Redbook sounded much better and clearer, in my opinion. The SACD sounded mildly muffled and distorted in comparison, very subtly, mind you, but noticable. It was really obvious in the hi hat cymbles. The Redbook had a nice "tss tss" where the SACD was more "csh csh" and at a lower pitch (??).

The SACD had a noticably lower overall level; I thought young dude turned the volume down at first, but he swears he didn't touch anything. I could swear I heard the harmonic distortion characteristic of 1-bit, not just around the edges, but all the way through the mix. It was in the high range, however, where the difference was most obvious. The Redbook had details and nuances that were dampened in SACD.

The SACD had a little bit less dimension than the Redbook.

The Klipsch speakers sounded vry nice. I might get myself a set for the fouton room.

That's my report. The Curve will prepare himself some dinner now, and watch The News Hour with Jim Lehrer. Bye!

E \:\)


Eric Vincent (ASCAP)
http://www.curvedominant.com
Joined: Jul 2001
Posts: 709
A
Gold Member
Offline
Gold Member
A
Joined: Jul 2001
Posts: 709
In the magazine TAS (The Absolute Sound) the editor wrote in an article with the headline: The Truth about SACD ".... it is probably worth noting, at this point, that the DVD-A disc has the support of most of the major labels save Sony, Telarc, and fistful a minor labels, but so far there has been anything but unanimity (thatis, a standard) for the sounds coming of those disc.
And that leaves us with SACD, a differnetial PCM system (called, by the technofroids, delta-sigma), using one bit (more in actual practice) and an out-of -this-world sampling rate (more than 2 megahertz). To make this beastie work, there must be all sorts of noise-shaping applied to the signal, virtually all of it above the range of contemporary electronics.
What Iīm saying, in short, is that the ultra-sonic, and quite noisy, SACD algorithms can (and to my ears, do) create perturbations further down, within the range of human hearing. (I think this is the source of the noisy artifacts I hear on the forte massed strings and brass in Telarcīs recent recordings of "Scheherazade" and Berliozī"Symphonie Fantastique". Knowing well that there might be first-generation problems with SACD technology, and wanting to see SACD at its best, I asked to audition the two-channel Accuphase DP-100 digital deck and its DC-101 processor, which received a rave review in these pages and which costs some $ 28,000 dollars.
..... Beyond that, in direct A/B comparisons with the Burmester CD gear on Telarcīs dual layer CD-compatible SACD discs, the Accuphase came out a distant (read: Tibetan) second (Burmester 969, the playback deck, comes in at $ 27,930, and the 970, its upsamplig DAC, at $ 30,670).
The two-piece Accuphase set was colored in the same way one would hear if he crossed an old Conrad-Johnson tubed unit with something from Sonic Frontiers; that is, it sounded warm, a bit lush - especially mid-bass - and veiled, as if a gauzy scrim had been droped between the reproduced sound and the listener. The one area of superiority lay in the Accuphaseīs reproduction of depth and space, characterisics we might have attributed to the fact that we were listening to two-channel versions of surround-sound originals. There was none of the Burmesterīs exquisitely detailed high frequencies, none of its tonal neutrality, little of its wide dynamics, and none of its rock-em, sock-em deep bass. In other words, playing regular old 16/44 CDs, the Burmester shamed the Accuphase to a degree I would not have thought possible."

This quotation is from TAS issue 134, page 29.
I think we donīt need DSD for high resolution audio, maybe for cheap consumer players with cheap DA converters.

Joined: Jan 2002
Posts: 13,114
D
10k Club
Offline
10k Club
D
Joined: Jan 2002
Posts: 13,114
Great thread. I don't understand a word of it, because I have no clue what DSD, dCS, and FPGA mean. For all I know -

- DSD is a popular new hallucinogen

- dCS is a recently-approved abortion procedure

- FPGA is a golf league for fat people

Judging from some of the posts I've read, I don't think I'm the only one who's confused. Could someone post a concise primer - PLEASE?


The Black Knight always triumphs!
Joined: Jan 2000
Posts: 5,160
Likes: 23
MP Hall of Fame Member
Offline
MP Hall of Fame Member
Joined: Jan 2000
Posts: 5,160
Likes: 23
Quote:
Originally posted by Dan South:
Great thread. I don't understand a word of it, because I have no clue what DSD, dCS, and FPGA mean.


Apparently it would mean nothing unless you have a stereo made by Manley that costs as much as the house I live in.

I think the primary thing you need to know here is the concept of a clash in the philosophy of increased signal resolution at a limited sample frequency versus an extreamly limited signal resolution at a ridiculous sample frequency - Direct Stream Digital. Right?

Personally, I don't know anyone who even has a DVD-A player, and only a few could explain what one is, much less a DSD setup....


Hmm.. Thinking... Isn't all of this out of context relative to the way speaker drivers work?

One would presume a high frequency driver is a non-perfect device, in that it's going to be "missing" harmonic detail in certain events captured by a certain sample rate, played back at a certain sample rate even in a perfect electrical amplification system. In other words, if one sample goes to digital nothing for 1 clock cycle and then back to full level, at 96K it's probably not going to "catch"/reproduce that accurately - whether it's periodic or not. In which case the driver would probably be more influenced by a cumulative "mechanical rounding off" of the signal, based on perhaps the average of what happens across 3 samples instead of 1.

Hmm. Really, drivers sort of act as a mechanical "frequency dither"; they're going to round off events that happen at ultra high sampling playback frequencies. They don't reproduce them, they just miss them or turn them into harmonic distortion if they can't react fast enough, right? They *can* respond with an infinite analog resolution within their operating bandwidth.

In which case the more accurate the sample is across that error range, the more accurate the reproduction will be. Increase clock frequency and I think you're getting diminishing returns mechanically since a driver isn't going to be able to keep up-ballistics wise, right? On the other hand, increased resolution of bit depth may yield a finer interaction with the mechanical dynamic of the driver since a minute averaged but more accurate fluctuation over a longer time span may yield a
better mechanical reproduction instead of less bit depth at a faster rate - 10 "blurry" samples in the same time span as 3 "detailed" samples. The 3 detailed samples would translate better than a whole bunch of less detailed samples that the driver can't discern routinely anyhow.

So ultra-high/higher sampling frequency without taking the mechanical means of reproduction into account doesn't make sense, while higher bit depth could translate more accurately since mechanical drivers are capable of "infinite" resolution within their operating bandwidth.

Right?

Man, I love making posts half asleep, always humbling when awake...


Guitar Lessons in Augusta Georgia: www.chipmcdonald.com
Eccentric blog: https://chipmcdonaldblog.blogspot.com/

/ "big ass windbag" - Bruce Swedien
Joined: Oct 2000
Posts: 4,223
C
MP Hall of Fame Member
OP Offline
MP Hall of Fame Member
C
Joined: Oct 2000
Posts: 4,223
Quote:
What Iīm saying, in short, is that the ultra-sonic, and quite noisy, SACD algorithms can (and to my ears, do) create perturbations further down, within the range of human hearing. (I think this is the source of the noisy artifacts I hear on the forte massed strings and brass in Telarcīs recent recordings of "Scheherazade" and Berliozī"Symphonie Fantastique". Knowing well that there might be first-generation problems with SACD technology, and wanting to see SACD at its best, I asked to audition the two-channel Accuphase DP-100 digital deck and its DC-101 processor, which received a rave review in these pages and which costs some $ 28,000 dollars.

..... Beyond that, in direct A/B comparisons with the Burmester CD gear on Telarcīs dual layer CD-compatible SACD discs, the Accuphase came out a distant (read: Tibetan) second (Burmester 969, the playback deck, comes in at $ 27,930, and the 970, its upsamplig DAC, at $ 30,670).
The two-piece Accuphase set was colored in the same way one would hear if he crossed an old Conrad-Johnson tubed unit with something from Sonic Frontiers; that is, it sounded warm, a bit lush - especially mid-bass - and veiled, as if a gauzy scrim had been droped between the reproduced sound and the listener. The one area of superiority lay in the Accuphaseīs reproduction of depth and space, characterisics we might have attributed to the fact that we were listening to two-channel versions of surround-sound originals. There was none of the Burmesterīs exquisitely detailed high frequencies, none of its tonal neutrality, little of its wide dynamics, and none of its rock-em, sock-em deep bass. In other words, playing regular old 16/44 CDs, the Burmester shamed the Accuphase to a degree I would not have thought possible."

This quotation is from TAS issue 134, page 29.
This description of SACD is an even more articulate, yet totally accurate, description than the one I described, in what was as far as I could ascertain a completely controlled A/B test. The SACD sample sounded noticably worse than the redbook 16 bit 44.1K sample, without question. "...as if a gauzy scrim had been droped between the reproduced sound and the listener" is exactly correct! SACD veils the sound of the overall mix with a completely artificial coloration that is not pleasing in any way, and it cannot be taken out at all, by any means.

This demonstrates the power (and beauty) of science, and math: you cannot argue with hard math. How do you properly dither a 1-bit signal? You cannot, because you need two dithering bits (white noise with a triangular pdf and a width of 2 LSBs), plus 1 bit audio, so you need at least a three-bit depth in order to properly dither, or you get a shitty signal. Why 2 dithering bits? Because triangular probability noise has an average energy value of only 0.5 of the peak value, so you need twice the level range than 1bit to pass it correctly without clipping (remember that clipping it will destroy the triangluar PDF). So at least 2 bits are required for the dither, so practically speaking you need 3bits to pass dither and some audio to get no distortion from the data channel. Mathematically, there is simply no way around that. SACD is 1-bit. How does it get around the math? It doesn't, because it cannot, and that is why it sounds inferior.

Over and out...

Eric Vincent
Curve Dominant Sound&Vision
curvdominant@earthlink.net


Eric Vincent (ASCAP)
http://www.curvedominant.com
Joined: Oct 2001
Posts: 63
P
Senior Member
Offline
Senior Member
P
Joined: Oct 2001
Posts: 63
Quote:
Originally posted by Curve Dominant:

you cannot argue with hard math


I can't having studied math for two years and let that pass !

This sentence is too much used to provide an alibi to an unproved assertion. Math only say what you make them say.

So speaking of math, could you first explain what you call "dither" for a DSD signal ?

I thought that dithering was a process in which the samples values of a PCM signal were modified adding them a random value. So what is dither for a DSD signal, that have no sample values ?

Joined: May 2000
Posts: 14,215
Cosmic Cowboy
10k Club
Offline
Cosmic Cowboy
10k Club
Joined: May 2000
Posts: 14,215
Wow...I got to page four and all of a sudden I had a splitting headache...this could be worse than Nika's "...96K" thread... ;\)

At this point I consider my self an "outside observer" of all the SACD/DSD technology. But I did read a very interesting opinion in the USA February issue of AudioMedia Magazine. It can be found in the magazine's column toward the back called "The Last Word...".

I tried to link to it on the AudioMedia website...but they have not yet archived that issue.
So, I have taken the liberty of copying it word for word and posting it here, assuming that since it was printed in a magazine it was intended for mass consumption.

It might help or hinder some of the opinions expressed thus far.

This was submitted to AudioMedia by:
Ron Rolland
Fine Audio Recording Services
Lake Forest, Ill.

Quote:
"The Last Word" column from AudioMedia Magazine Feb 02, by Ron Rolland:

Have We Been Blinded By Science?

I think it is high time that someone give an honest appraisal of Sony's SACD/DSD format. First of all, as a professional recording technique, DSD is incompatible with every PCM digital audio device developed thus far. It requires proprietary equipment just to record the signal, with editing presenting a whole other collection of problems. Once again, the brain trust at Sony has produced another branded format, not for the questionable sonic improvement it may offer - but for the creation of a rich marketplace of software and licensed hardware, a market they feel they can drive down our throats, as owners of the largest library of recorded music in the world (Columbia).

In an effort to hear the actual sonic quality of SACD/DSD (beyond my own Genex/DSD Stereo recordings), I acquired a Sony SCD-777ES player. The problem is that you can only derive SACD playback from the line-level outputs, which feed through Sony O/P amplifiers notorious for their bright quality. So, how do you actually hear SACD/DSD to make any honest comparison? In our studios we use dcs 904 and 954 AD/DACs that can encode and decode DSD digital audio when played back from a Genex or Sony recorder. However, there are no facilities to playback SACD from a disc. After some investigation, I discovered that the only way to hear SACD by way of digital input to a high quality DAC is to use the DCS Verdi SACD Transport with an 1394EEE firewire link into the dcs Elgar DAC. That's all well and good except you must shell out $9,000 for the transport and $10,000 for the DAC. I would say most consumers might find this gear a little out of reach, especially to hear re-masters of the Columbia library that they already have in their collections. There are a handful of excellent original recordings by Jack Renner and Michael Bishop of Telarc and Tom Jung of DMP, but that's a lot of dough to hear these few.

So what does the average consumer experience with the new Super Audio CD? A molded plastic chassis housing mass produced chips and cheap output amplifiers … for $299. The quality of the analog components alone is so poor that it is difficult to discern any difference in presentation between a basic CD and the Super pretender. SO, what is the point? It is through no quest for advancing the fidelity of recordings purchased by the consumer that's for sure. It is, as usual, greed that drives this whole effort. Even though there are serious criticisms regarding overloading in a 1bit system, I do agree, DSD has been and can continue to be an excellent archival format for audio. It could be applied to studio databases for central library storage, but thus far only the audio erudites have heard and can afford this new technique and they have been largely facilitated by Sony to advance the format.

(AudioMedia) Editor's Note: Sony was contacted but declined to comment on this letter.
Enjoy! \:D


miroslav - miroslavmusic.com

"Just because it happened to you, it doesn't mean it's important."
Page 6 of 6 1 2 3 4 5 6

Moderated by  gm 

Link Copied to Clipboard
Powered by UBB.threads™ PHP Forum Software 7.7.5