Linear PCM is crap, SACD is best

Topic: Linear PCM is crap, SACD is best (Read 37379 times) previous topic - next topic

0 Members and 1 Guest are viewing this topic.

Linear PCM is crap, SACD is best

2004-03-30 12:26:26

[span style='font-size:7pt;line-height:100%']This is split from Is DVD Audio necessary ? by tigre
---------------------------------------------------------------------------[/span]

Dear readers,

Apart from comments about mastering and producing the audio material, there is a lot of rubbish in the posts...

If I am not mistaken, the question was: do we need DVD-Audio or SACD. Well, the answer is very simpel. We don't need DVD-Audio, but we need SACD instead.

To give answer to all those non-audio guys out there: YES, YOU CAN HEAR THE DIFFERENCE BETWEEN CD AND SACD!!! In fact, you can hear it immediatly.

The reason for DVD-Audio to fail is also very simpel: linear PCM. Those guys who "invented" this format are living in the past and by now studies show that 24 bits are indeed not enough. The problem is that the information needed for pinpointing the music is at frequencies at which the levels are at -40 dBFS. This means 7 bits right to the recycle bin.

Another explain for non-audio guys: yes, you need more bandwidth that 20 kHz. Not that we can hear it, but it has its impact for filtering. However, perception studies show that we have detection hairs in our ear to detect 50 kHz! Maybe we can't hear it as such, but it is detected and is doing something...

Last but not least, one can mathematical prove that SACD is the best by means of impulsresponses. By the way, high-end music material is recorded and stored by studios in SACD-like format. Here every level has the more or less the same resolution thanks to the 1-bit class D type of sampling (DSD). All other formats, including DVD_Audio, are derived from this. So why downsampling what is already there: SACD

So, please throw DVD-Audio out of the window and spread the word.

Jacco

Linear PCM is crap, SACD is best

Reply #1 – 2004-03-30 12:58:11

Quote

Dear readers,

Apart from comments about mastering and producing the audio material, there is a lot of rubbish in the posts...

If I am not mistaken, the question was: do we need DVD-Audio or SACD. Well, the answer is very simpel. We don't need DVD-Audio, but we need SACD instead.

To give answer to all those non-audio guys out there: YES, YOU CAN HEAR THE DIFFERENCE BETWEEN CD AND SACD!!! In fact, you can hear it immediatly.

The reason for DVD-Audio to fail is also very simpel: linear PCM. Those guys who "invented" this format are living in the past and by now studies show that 24 bits are indeed not enough. The problem is that the information needed for pinpointing the music is at frequencies at which the levels are at -40 dBFS. This means 7 bits right to the recycle bin.

Another explain for non-audio guys: yes, you need more bandwidth that 20 kHz. Not that we can hear it, but it has its impact for filtering. However, perception studies show that we have detection hairs in our ear to detect 50 kHz! Maybe we can't hear it as such, but it is detected and is doing something...

Last but not least, one can mathematical prove that SACD is the best by means of impulsresponses. By the way, high-end music material is recorded and stored by studios in SACD-like format. Here every level has the more or less the same resolution thanks to the 1-bit class D type of sampling (DSD). All other formats, including DVD_Audio, are derived from this. So why downsampling what is already there: SACD

So, please throw DVD-Audio out of the window and spread the word.

<rant>

Before you come here preaching at people, please see just how many times this has already been discussed in the FAQ.

EDIT: http://www.hydrogenaudio.org/forums/index....7516#entry74075

</rant>

Cheers,
David.

Linear PCM is crap, SACD is best

Reply #2 – 2004-03-30 13:04:54

dekkersj, you've been HA member for almost 4 months now, so it's hard to believe that this is a newbee mistake. Your post is one of the most obvious TOS # 8 violations I've seen here for a long time.

Quote

YES, YOU CAN HEAR THE DIFFERENCE BETWEEN CD AND SACD!!! In fact, you can hear it immediatly.

Provide evidence for this, please.

Quote

studies show that 24 bits are indeed not enough. The problem is that the information needed for pinpointing the music is at frequencies at which the levels are at -40 dBFS. This means 7 bits right to the recycle bin.

What studies? Links and quotes, please. In fact the quantization noise of 24bit resolution is lower than the noise floor caused by thermal molecule movement in analog components.

Quote

However, perception studies show that we have detection hairs in our ear to detect 50 kHz!

1. I doubt that this is true. Please tell the source of this information
2. The human body can react to even higher frequencies (at high 'volume'), e.g. the ultra sound used for smashing kidney stones. This doesn't mean it's audible.

Quote

Last but not least, one can mathematical prove that SACD is the best by means of impulsresponses.

So what? Show me a listening test that proves that CD resolution isn't already enough for transparent reproduction of impulses.

Quote

By the way, high-end music material is recorded and stored by studios in SACD-like format.

Great argument. This means you define all music material stored in another way by studios as "not high-end"

Linear PCM is crap, SACD is best

Reply #3 – 2004-03-30 13:18:47

I don't preach. I tell you how it is.

In my work, we have the privilege to "create" those discussions David is referring to. Also to verify the outcome.

The reason of being so hard and harsh is that there are a lot of commercial things going on, especially in the field of DVD-Audio. Never believe those selling-guys!

The same sales-boys were trying to sell a 14 (yes, 14 and not 16) bit CD-player in 1983, with the claim that is was perfect. Not!

It is really breathtaking to listen to a real multichannel SACD, you will never go back to CD anymore.

greetz,
Jacco

Linear PCM is crap, SACD is best

Reply #4 – 2004-03-30 13:53:27

Dear tigre,

It is about time someone is going to clarify things and that is always a dirty job.

Anyway,

How can I prove to you that you can hear the difference? Let's turn it around: if you don't hear it, don't buy it and go to see a doctor. In my line of work and hobby I have seen dozens of people that could double-blind tell me what a CD is and what a SACD. Never done this experiment with DVD-Audio. More interesting would be a test with SACD and DVD-Audio. Everybody with good ears can tell you the difference between CD and SACD, absolutely no question about that!

Those studies that 24 bits are not enough were done recently (not by us) and can be found on the internet. If I have some time I will look it up for you. It has nothing to do with noise levels or whatsoever, just with resolution reduction caused by the linear PCM principle. Throw it away I would say...(maybe another TOS violation, I don't care it is the truth).

The perception studie is done at Delft University and in Eindhoven there is also work in this field. Here, the length of those hairs is corresponding those high frequencies! I never said it is audible, I meant it is detectable.

The best way to demonstrate the inability of CD to reproduce impulses is to listen to music and compare it to SACD. This is what you have to do yourself and never trust on conclusion of others (including me). After doing that, you will admit that SACD will resemble the truth more closely that anything you have ever heard before (exept for the original in the concert hall of course). Or else make an appointment with your doctor.

And yes, with the arrival of SACD, all other formats became not high-end. Sharp and quick conclusion.

By the way, I have these discussions day in day out with all kinds of people, but they never give me obvious TOS # 8 violations.

Linear PCM is crap, SACD is best

Reply #5 – 2004-03-30 14:04:44

Quote

[snip]

Is this wrong?

almost entirely

There's an element of truth in it, but the idea that "the quiet sounds are recorded at 8-bit resolution, and we all know how bad that sounds" is rubbish. It doesn't matter how many bits you're "using" at that moment, a 16-bit recording allows a noise floor 90dB (100dB + with noise shaped dither) below the peak.

If you twiddled your volume control so that the peak signal is 100dB SPL in your living room, that means the noise floor is at 10dB, (orunder 0dB SPL with noise shaped dither). Unless you live in a sound proof room, other sounds will hide this! Anything below 0dB SPL is inaudible, even in complete silence.

More or less bits are doing nothing magic - they just move the noise floor up and down relative to digital full scale. With correct dither, the system is entirely linear, and sounds are present well below the noise floor.

In my previous post, I included a link to the excellent FAQ, which in turn links to some very excellent threads!

Cheers,
David.

P.S. "almost" entirely, because to preserve the correct shape of the largest (not painful) waveform peak, and ensure the noise floor is below 0dB SPL, you do need more bits. There are few recordings which need this!
P.P.S. "almost" entirely, because "the system is entirely linear" only to the second order - dither theory shows the error still has some trace of the signal in it, if you know how to look. No one has ever suggested this as a source of audible error.

EDIT: I wrote this response immediately after my previous one, but the board crashed!

Linear PCM is crap, SACD is best

Reply #6 – 2004-03-30 14:15:38

Quote

Quote
Last but not least, one can mathematical prove that SACD is the best by means of impulsresponses.
So what? Show me a listening test that proves that CD resolution isn't already enough for transparent reproduction of impulses.

This is wrong too ! This is at best a misunderstanding, and possibly a dishonest marketing trick.
I just explained why in the thread about super tweeters :

http://www.hydrogenaudio.org/forums/index....ndpost&p=198057

What magazines and technical documentation call "better impulse response" are nothing else than the same impulses as the "bad" ones, but with ultrasonic content.

I admit, that this is not mathematically obvious, one has to know that a perfect pulse is a Dirac distribution, that the the Fourier transform of a Dirac distrubution is a constant function, that the inverse Fourier Transform of a step function (lowpass of the previous) is a cardinal sinus, and that a cardinal sinus looks like a bad impulse response !

But while it looks like a bad impulse response, it is not. It is an ideal impulse response under a given lowpass !

Linear PCM is crap, SACD is best

Reply #7 – 2004-03-30 14:19:26

Quote

I don't preach. I tell you how it is.

In my work, we have the privilege to "create" those discussions David is referring to. Also to verify the outcome.

The reason of being so hard and harsh is that there are a lot of commercial things going on, especially in the field of DVD-Audio. Never believe those selling-guys!

The same sales-boys were trying to sell a 14 (yes, 14 and not 16) bit CD-player in 1983, with the claim that is was perfect. Not!

It is really breathtaking to listen to a real multichannel SACD, you will never go back to CD anymore.

greetz,
Jacco

Whoa, Jacco. Talk about cruising for a bruising.

As David said, that original post of yours was ill-advised even for a newbie, let alone a member of 4 months standing.

" ..... Never believe those selling-guys!"

Seems to me you've bought the Sony-Phillips marketing schpiel hook,line and sinker.

If you want a definitive analysis of what's going on within DSD encoders/decoders, try this;

why DSD/1-bit sigma-delta is unsuitable for high-quality applications

ciao,
RF

Linear PCM is crap, SACD is best

Reply #8 – 2004-03-30 14:22:08

Quote

It is really breathtaking to listen to a real multichannel SACD, you will never go back to CD anymore.

Oh, I don't know - I've heard some really terrible multi-channel SACD demos from Sony. Some good ones too.

Ditto DVD-A. Anyway - multichannel isn't the point (though it's a bigger selling point than greater than 44.1kHz sampling and greater than 16-bit quantisation!)

(I'll regret asking this, but...) What's your professional experience of SACD then?

Did you make it through all those threads in the FAQ? I can't believe you even had time to read them between posts!

For example...

Quote

Here every level has the more or less the same resolution thanks to the 1-bit class D type of sampling (DSD). All other formats, including DVD_Audio, are derived from this. So why downsampling what is already there: SACD

"Class D" is a slightly inappropriate label, since analogue class D PWM technology doesn't have a fixed "sample point".

More importantly, though some studios use DSD-wide for editing (64fs, 8-bit), most use 24-bit linear PCM, either at 48kHz or 96kHz. As for DVD-A being derived from SACD - you can convert either way, but you'll find many SACDs are taken from 24/96 masters (or worse!) simply because the majority of people recording and master to multibit.

Quote

In my work, we have the privilege to "create" those discussions David is referring to. Also to verify the outcome.

Your assumption that you should come to HA to give us the kind benefit of your professional knowledge, presumably because we're all a bunch of bumbling amateurs who couldn't earn a living with our audio knowledge and don't know what we're talking about, will raise a smile or two.

Anyway, I humbly refer you to the FAQ.

Cheers,
David.

Linear PCM is crap, SACD is best

Reply #9 – 2004-03-30 14:31:53

Quote

Quote

Quote
Last but not least, one can mathematical prove that SACD is the best by means of impulsresponses.
So what? Show me a listening test that proves that CD resolution isn't already enough for transparent reproduction of impulses.

This is wrong too !

Why? If you take a dirac impuse at very high sampling rate, e.g. 192kHz and convert the signal to 44.1kHz (in a decent way), the result won't be silence - single sample click - silence anymore, but something like in the graphs you linked to - true. But I doubt that there's an audible difference, that's what I meant. (Or was your comment aimed at the quote in the quote?)

Linear PCM is crap, SACD is best

Reply #10 – 2004-03-30 14:37:18

Quote

but the idea that "the quiet sounds are recorded at 8-bit resolution, and we all know how bad that sounds" is rubbish.

Well, well, well...The point is that linear PCM is rubbish. Of course is it a matter of bits you are "using" at the moment. This is called resolution. The better the resolution, the better the music will be (to whom am I this explaining anyway...). Please note that a noise floor has nothing to do with it.

cheers,
Jacco

Linear PCM is crap, SACD is best

Reply #11 – 2004-03-30 14:58:17

Quote

(I'll regret asking this, but...) What's your professional experience of SACD then?

Don't feel sorry, I work as a research engineer with Philips and I can assure you: never believe the sales-boys:

Quote

Seems to me you've bought the Sony-Phillips marketing schpiel hook,line and sinker.

I have some Philips material, altough modified of coarse.

We developed both the CD and the SACD on this floor and performed a lot of listening test with material how it should be. In respond to

Quote

Oh, I don't know - I've heard some really terrible multi-channel SACD demos from Sony. Some good ones too.

Ditto DVD-A. Anyway - multichannel isn't the point (though it's a bigger selling point than greater than 44.1kHz sampling and greater than 16-bit quantisation!)

, I can only say: too bad you picked the wrong examples. It is commercial so everybody will press his material onto a SACD in the hope never will hear that it is a bad recording...

Obviously, this has to change:

Quote

most use 24-bit linear PCM, either at 48kHz or 96kHz.

But probably, we all will cry in a few years since the majority of the people is happy enough with dolby digital and/or dts shit...

The only thing what we don't understand over here, is why oversampling an existing digital stream results in better sound. Up to now there is no mathematical prove found. Challenge in other words.

cheers,
Jacco

Linear PCM is crap, SACD is best

Reply #12 – 2004-03-30 15:08:37

Quote

Which means that the extra 8 bits achieve nothing. Better DACs are certainly available but unless you are using very high end equipment, 16 bit is just as good as 24 bit.

The trick is the fact that the arithmetic part of such an IC is 24 bits long. For normal reproduction is more bits not really needed, but when there is processing of the data (interpolation or volume control or whatever) more bits are better. Less errors.

An interesting subject:

Quote

I have still not seen concrete evidence to prove that a 22kHz brickwall filter does not change our perception of music.

Here in our lab we did SACD-experiments with people with "golden ears" so to say. A very strange one was an audio piece with external low-pass filtering. The cutoff was set to 50 kHz and the question was: can you hear the filter? The answer was Yes, they were able to identify the situation with or without filter. The most strange experiment was added filtering on the music data itself with a linear phase filter. So no group delay and still those guys were able to identify the original and the manipulated data. I have no public reports on this, but you have to believe me. So brickwall filtering can be assumed as desastrous...

cheers,
Jacco

Linear PCM is crap, SACD is best

Reply #13 – 2004-03-30 21:12:27

Quote

was your comment aimed at the quote in the quote?

Yes, it is the quote in the quote that sounds wrong.
Sorry for the confusion.

Linear PCM is crap, SACD is best

Reply #14 – 2004-03-30 22:09:58

Quote

The trick is the fact that the arithmetic part of such an IC is 24 bits long. For normal reproduction is more bits not really needed, but when there is processing of the data (interpolation or volume control or whatever) more bits are better. Less errors.

Of course. My point was that for the standard consumer transport->dac->analog volume control->amplifier chain, 24 bit LPCM has no advantage over 16 bit.

Quote

Here in our lab we did SACD-experiments with people with "golden ears" so to say.

These tests and their conclusions sound like that kind of thing that people in this community would be very interested in. Could you try and get the reports released to the public. If these reports are as conclusive as you imply then they might have a very positive effect on Phillip's marketing. Communities like this one like nothing more than a good technical report. Failing getting the actual reports released, you at least share the details of how these tests were conducted so we could attempt to re-create the results you quote.

Without reports or our own results we have a very hard time believing the claims you make which go against the experiences of many board members.

Linear PCM is crap, SACD is best

Reply #15 – 2004-03-30 22:44:52

Quote

[span style='font-size:7pt;line-height:100%']This is split from Is DVD Audio necessary ? by tigre
---------------------------------------------------------------------------[/span]

Dear readers,

Apart from comments about mastering and producing the audio material, there is a lot of rubbish in the posts...

If I am not mistaken, the question was: do we need DVD-Audio or SACD. Well, the answer is very simpel. We don't need DVD-Audio, but we need SACD instead.

To give answer to all those non-audio guys out there: YES, YOU CAN HEAR THE DIFFERENCE BETWEEN CD AND SACD!!! In fact, you can hear it immediatly.

The reason for DVD-Audio to fail is also very simpel: linear PCM. Those guys who "invented" this format are living in the past and by now studies show that 24 bits are indeed not enough. The problem is that the information needed for pinpointing the music is at frequencies at which the levels are at -40 dBFS. This means 7 bits right to the recycle bin.

Another explain for non-audio guys: yes, you need more bandwidth that 20 kHz. Not that we can hear it, but it has its impact for filtering. However, perception studies show that we have detection hairs in our ear to detect 50 kHz! Maybe we can't hear it as such, but it is detected and is doing something...

Last but not least, one can mathematical prove that SACD is the best by means of impulsresponses. By the way, high-end music material is recorded and stored by studios in SACD-like format. Here every level has the more or less the same resolution thanks to the 1-bit class D type of sampling (DSD). All other formats, including DVD_Audio, are derived from this. So why downsampling what is already there: SACD

So, please throw DVD-Audio out of the window and spread the word.

Jacco

Jacco,
Nice troll.

You make claims with no hard evidence to back it up. Some pointed questions for you:

1) Technically DSD has issues that can make it inferior to LPCM; as a matter of fact CD has a higher dynamic range from 10kHz to 20kHz than DSD. What technical argument do you have that DSD is a superior encoding technology to LPCM? LPCM is very well behaved in both the frequency and the amplitude domains and very easy to process. DSD is not, although it can still be a very good format.

2) Audible effects of LPF below 50kHz? Show evidence or let it go - very old topic with inconclusive results.

3) 24 bits of resolution is not enough? Show me any piece of audio equipment with a dynamic range greater than 24 bits (144dB) or let this go. I can buy needing more than 24 bits during processing, but not at the consumer level.

4) My understanding is that the vast majority of studio recordings are LPCM for much of the recording process and that DSD is a post-processing step for most recordings, not the other way around. Most recording consoles and HW/SW don't support DSD from what I have read - perhaps you have factual evidence otherwise?

5) Search for a thread about Pink Floyd's Dark Side of the Moon SACD vs. CD - the mastering was messed up for the CD layer meaning that you can't directly compare the two technologies on this disc, or many others I suspect. So how can you begin to even evaluate the DSD vs. LPCM topic in a consumer format?

6) You can mathematically prove that DSD has a superior impulse response to LPCM? Show me.

7) SACD and DVD-Audio are crippled formats with copy protection, encryption, watermarking, no PC support, extremely limited input/output support (6 analog patch cords?). Why do we need SACD?

So the bottom line is: SACD vs. DVD-Audio is a "so what" topic. They are both dinosaurs of the 1990s music industry thinking. Show that DSD is superior to LPCM per your dubious claims or drop the subject. I submit that even if DSD (SACD) is superior to LPCM it still doesn't matter because SACD is a crippled format and many modern recordings are so compressed/clipped that any difference in sound quality is more limited by the master than by the format. The answer to the SACD vs. DVD-Audio question is: CD.

Linear PCM is crap, SACD is best

Reply #16 – 2004-03-31 09:45:23

Quote

Jacco,
Nice troll.

He wasn't trolling. He was promoting his professional interest.

I was hoping he was going to tell us if Philips have done any tests which people on this board might find relevant. But it seems we've scared him off!

If not, dekkersj, what double-blind tests have been done on SACD vs CD or DVD-A, and do Philips have any technical explanations for the claimed superiority of SACD?

Do you know of any published papers (peer reviewed, or otherwise) reporting double blind listening tests that demonstrate that an ultrasonic low pass filter is audible?

EDIT: To everyone else - you ask for ABX proof of SACD vs CD, but it's well known that audio equipment sounds different when you add high levels of ultrasonic noise to the input signal. The real question is, does SACD sound different from CD+similar ultrasonic noise?

Cheers,
David.

Linear PCM is crap, SACD is best

Reply #17 – 2004-03-31 10:46:22

Hello everybody,

I received an e-mail from David with the question to respond. Well, I didn't realize that there was that much posts regarding this topic. There were a lot of questions and I will try to answer them as many as possible.

To start with a commercial story (which I don't want to preach because I hate commercial things), one can state that SACD has began as a succesor for the CD. Patent-wise and Sony/Philips were looking for a new cash cow. Research on the topic DSD is very old and it was intended for backing up material on a medium that was so bad that after one time of playing the tape was destroyed. The other reason of developing SACD originates from the recording industry. They wanted to have a succesor of the CD with 12 constraints. Copy protection and so on.

Quote

If these reports are as conclusive as you imply then they might have a very positive effect on Phillip's marketing.

Well what can I say, as far as my point of view is concerned: I don't understand the marketing of Philips and I hear rumors that the priority of SACD is set low and the research (where I am working) is stopped. The last thing we solved was a timing problem in SA mixing tables of Sony. By the way, Research and Consumer Electronics are two companies within one large company and they don't listen that much to each other...

Another interesting topic according to the topics is Dynamic Range. I measured myself a common amplifier and in normal listening level it has a signal-to-noise ratio of 75 dB. I also measured good tube amps with a SNR of 85 dB. IMHO it has more to do with resolution and there is were DSD comes in. The statement

Quote

as a matter of fact CD has a higher dynamic range from 10kHz to 20kHz than DSD

is true but irrelevant. When the CD was brought to the world, audio enthousiasts rejected it quit soon and said:"CD is for boom-boom music..." A record player (with a good element of coarse) sounded much better, ie. with extra noise, scratches and so on but at the end there was a better soundpicture than with CD. It still is.

It is more important to have large resolution in the low levels and no low pass filtering. (needed for CD)

Quote

and do Philips have any technical explanations for the claimed superiority of SACD

The low pass filtering is considered as the best explanation of why the sound is that good. (records were made with 50 kHz low pass filtering)

Quote

Do you know of any published papers (peer reviewed, or otherwise) reporting double blind listening tests that demonstrate that an ultrasonic low pass filter is audible?

The outcome of this test intriges me most at the moment, but it is still confidential since we don't understand what is going on. I don't want to make promisses which I can't make true, regarding to a report publication. There are strange things to be explained and this one of the strangest.

Quote

Pink Floyd's Dark Side of the Moon SACD

is a bad example since it is not a SACD recording. If you are used to SACD (when I refer to SACD, I mean a real SACD recording not a upconverted 24/96 signal...) you can distinguish such an old recording as with Pink Floyds directly from a SACD recording. Also in the recording process there are differences where the most important thing is phase coherence between the channels.

I admit, DSD has many artifacts such as processing, HF noise (we blew up a very expensive amplifier once with this) but the end result to the ear is most important. It is also the format which complies to every demand in the recording industry, and is therefore for them the best candidate. However, some audio crunchers are busy with watermarking in the audio itself and they must be stopped of coarse. When one listens to such a recording, one can sometimes hear ticks and cracks!

Quote

The answer to the SACD vs. DVD-Audio question is: CD.

If I have listened to a SACD and I have to go back to a CD, I always have disappointing feeling...

Linear PCM is crap, SACD is best

Reply #18 – 2004-03-31 11:37:13

Quote

To everyone else - you ask for ABX proof of SACD vs CD, but it's well known that audio equipment sounds different when you add high levels of ultrasonic noise to the input signal. The real question is, does SACD sound different from CD+similar ultrasonic noise?

I think it's still easy to perform a useful double-blind test if you have access to the necessary equipment (seems to be the case for dekkersj), similar to 24/96 vs. CD quality test:

Take a as-good-as-possible recorded/mastered DSD/SACD source (stereo).
-> Convert it to 16/44.1 with the best conversion algorithms (resampling/dithering) available. (Or - if you want to prove that SACD is superiour to DVD-A - convert to 24/96).
-> Convert back to DSD (source resolution).
-> Create a SACD with both, the original and the converted version
-> Let someone else play it back on best equipment available in random order and find out how often you can tell which version you hear.

This test will avoid many possible reasons for differences, e.g. equipment used.

Linear PCM is crap, SACD is best

Reply #19 – 2004-04-01 02:01:36

Quote

Take a as-good-as-possible recorded/mastered DSD/SACD source (stereo).
-> Convert it to 16/44.1 with the best conversion algorithms (resampling/dithering) available. (Or - if you want to prove that SACD is superiour to DVD-A - convert to 24/96).
-> Convert back to DSD (source resolution).
-> Create a SACD with both, the original and the converted version
-> Let someone else play it back on best equipment available in random order and find out how often you can tell which version you hear.

This test will avoid many possible reasons for differences, e.g. equipment used.

It'd be interesting to do something similar by mastering a CD from the DSD master, and then playing the SACD versus the CD + an ultrasonic white noise generator at a similar volume level to the ultrasonic component in the DSD recording (I am still curious whether the ultrasonic content needs to bear any relation to the normally audible content itself).

Quote

When the CD was brought to the world, audio enthousiasts rejected it quit soon and said:"CD is for boom-boom music..." A record player (with a good element of coarse) sounded much better, ie. with extra noise, scratches and so on but at the end there was a better soundpicture than with CD. It still is.

You might be interested in a portion of
this experiment, particularly the modified audio file uploaded by the indomitable PIO2001.

Linear PCM is crap, SACD is best

Reply #20 – 2004-04-01 02:42:21

Quote

Don't feel sorry, I work as a research engineer with Philips and I can assure you: never believe the sales-boys:

They hiring? Apparently I can just make my qualifications up on the spot

Quote

The low pass filtering is considered as the best explanation of why the sound is that good. (records were made with 50 kHz low pass filtering)

Let's not forget these threads sometimes come up when newbies are searching for answers. Misleading information on this board hurts us all.

Quote

maybe another TOS violation, I don't care it is the truth

Time for a banination?

Linear PCM is crap, SACD is best

Reply #21 – 2004-04-01 07:11:00

Jebus, thanks for trying to help to keep the forum's SNR high. You (and everyone else with similar ideas) might want to read this post and the following three posts from the thread where dekkersj made his claims originally. This thread here is still active because there's a chance that there'll be useful discussion and results. If this ends up as a mess, there are still ways to handle it.

Linear PCM is crap, SACD is best

Reply #22 – 2004-04-01 09:41:40

dekkersj,

You make a point of saying that 24bits is not enough, yet you never even tried to prove how many bits equivalent SACD has. Seeing as SACD has lower S/N ratio than 24/192 throughout its passband, and seeing as bit resolution directly defines the S/N ratio for any format where bit resolution has any direct significant meaning (and SACD is not one of them; if it were, its bit resolution is 1 and its S/N ratio would be somewhere around 6dB ), it would seem that SACD has an 'effective' bit resolution that is even lower than 24 bits.

And what does it mean for DVD-A if people really could tell the difference between full bandwidth and 50kHz lowpassed? DVD-A 24/192 records up to 96kHz, not 50kHz.

Linear PCM is crap, SACD is best

Reply #23 – 2004-04-01 16:59:45

I looked some things up regarding my claim that 24 bits is not enough, and it appeared to be on my laptop. An article from daisy-laser, if I am not mistaken they are one of the few companies that can produce SACD's. One page in particular is of interest and I made it readable here:
Why is DSD needed?

They put it very black and white by investigating a signal with rms level of -108 dBFS, but it shows clearly what I mean. It is clear that 24 bits is not enough to store a signal with such low values, although a lot of people seems to think otherwise. They believe that if the signal is above the noise floor of the quantizer, one can reconstruct the signal perfectly. This is of coarse not true at all...A linear PCM signal of 24 bits is at its best 24 bits of resolution since it is the peak. All other material is below or far below that 24 bits. This is the reason that many companies (and two large ones in particular) have concluded that multibit is not the way to go. (To put it black and white and in practise high-resolution Linear PCM and DSD have comparable audio quality)

By the way, I don't think one can compare the "equivalent bits" from SACD if one looks to SNR to the number of bits of a linear PCM system. (It is well known that here the normalized power from a rounding quantizer is q^2/12, where q=2^(-b) and b is the number of bits, resulting in SNR = 10log(12*2^(2b)*Pn), where Pn is the normalized power of the signal if all bits are 1) This is due to a different quantizing system for SACD. The sales department "claims" a more or less constant number of effective bits, ie. 24. I do not agree completely with this statement, but the bottom line is that for high frequency, yet audible, frequencies the resolution is far better than with Linear PCM. But this is easy to understand since smaller signals inherently have more distortion due to lower resolution, especially with normal cd. Only for boom-boom parts of the music, the 24 bits of LPCM are reached, the rest will have distorsion.

Quote

And what does it mean for DVD-A if people really could tell the difference between full bandwidth and 50kHz lowpassed? DVD-A 24/192 records up to 96kHz, not 50kHz.

That is very good for DVD-Audio and I am very happy with that. This subject was originally meant as an argument not to use low pass filtering at all and especially for brickwall filters and/or digital filters.

cheers,
Jacco

Linear PCM is crap, SACD is best

Reply #24 – 2004-04-01 17:46:07

Quote

One page in particular is of interest and I made it readable here:
Why is DSD needed?

That page doesn't show any more than that if you quantize to 24 bits a signal with a peak that's below 18 bits, the signal has an effective resolution of 6 bits.

This is so basic that I would be extremely surprised if any of the posters in this thread didn't already know it. (Except you, apparently)

It is saying that if you set your equipment so the lowest level recorded is at the threshold of audibility, then DVDA can store a sound loud enough to make you instantly deaf.

Or the reverse, if you set the maximum level of your equipment to a level that makes you instantly deaf, then the noise is at the threshold of audibility.

How again is this not sufficient?

Notice