Topic: DSD vs CD vs High-Res PCM (Read 14946 times)

Original_Ken · « **on:** December 14, 2013, 12:34:30 AM »

Yes, but it was a 13 year old article about ADCs and DSD. Kind of overall interesting from a historical perspective, but not really relevant much to DACs since they are barely touched upon.

Also, some of that text is misleading, because it makes it seem like DSD is crap, whereas despite all of the seemingly horrible noise, it still sounds better than CD.

And yeah, you did say "related perhaps"

Anaxilus · « **Reply #1 on:** December 14, 2013, 12:41:15 AM »

Quote from: Original_Ken on December 14, 2013, 12:34:30 AM

it still sounds better than CD.

Are you comparing the format or the mastering?

Marvey · « **Reply #2 on:** December 14, 2013, 01:11:09 AM »

DSD isn't crap. But DSD is actually BS as a high resolution format. The whole point of high resolution is to provide more information, more resolution, even it if this information supposedly outside what humans can hear.

The problem with DSD is that there's a lot of ultrasonic crap not too far up from the range of human hearing. This requires a filter to remove that crap otherwise your tweeters may blow. It's not uncommon for such a filter to start at 30kHz with a 24db/octave slop.

So much for hires. You if want a real high resolution format, better off with 24/96 PCM.

Original_Ken · « **Reply #3 on:** December 14, 2013, 01:16:50 AM »

Yes, I think that 24/192 is a better format, which is why it is supported by people like Barry Diament over DSD.

BTW, Jriver quietly added an option to MC19 to allow DSD to play with no filter. I tried it and I have yet to blow my headphones...

ultrabike · « **Reply #4 on:** December 14, 2013, 01:24:05 AM »

I also don't think DSD is crap at all. It just has it's limitations and restrictions like most things in life (TNSTAAFL) . In the end, like Anax said, there are many variables in comparing DSD to CD content. One of them being mastering, which sometimes is superior in even older formats like LPs.

Also note that old article doesn't really do the CD format any favors either given the 44.1 kHz brick-wall filter magnet sampling limitation, and 16-bit (which apparently almost became 14-bit) dynamic range cap. That comes with it's own set of issues as well.

Today however, on-line content may be available at higher sampling rates, free of the restrictions of the media format (like 24/96 PCM mentioned above). If this is the case, DACs that can exploit this fact maybe able to do a better job... That still doesn't say much about the mastering and recording which in many cases maybe the limiting factor.

Now 24/192 is a bit controversial though. The extra sampling rate might relax the analog filtering requirements, but too much of it might stress amp requirements.

Original_Ken · « **Reply #5 on:** December 14, 2013, 01:26:17 AM »

Not sure what you mean by "amp requirements" ? You don't mean analog amp ?

ultrabike · « **Reply #6 on:** December 14, 2013, 01:26:55 AM »

Yup. The one that the DAC feeds and drives the speakers/cans.

... Well, if the DAC analog LPF corner frequency and roll off is the same for both 96 and 192 it proly doesn't matter much.

Quote from: Original_Ken on December 14, 2013, 01:16:50 AM

Yes, I think that 24/192 is a better format, which is why it is supported by people like Barry Diament over DSD.

BTW, Jriver quietly added an option to MC19 to allow DSD to play with no filter. I tried it and I have yet to blow my headphones...

Sorry I missed this...

The ultrasonic stuff may not blow every single driver in existence, but it may stress some perhaps unnecessarily so... It may also stress amps.

So what you are saying is that playing 24/192 over DSD is a better format than 24/96 PCM... because?

Marvey · « **Reply #7 on:** December 14, 2013, 02:54:35 AM »

Quote from: Original_Ken on December 14, 2013, 01:16:50 AM

Yes, I think that 24/192 is a better format, which is why it is supported by people like Barry Diament over DSD.

BTW, Jriver quietly added an option to MC19 to allow DSD to play with no filter. I tried it and I have yet to blow my headphones...

24/96 is already a better format than DSD.

But try running the DSD->PCM conversion without a filter into an amp which is not necessarily stable at ultrasonic frequencies into tweeters and see what happens. Give it some time for for the voice coil to heat up too. Put it this way, I'm not willing to try this test with my Fostex T90 supertweeters. Although it's probably fine with headphones.

See pic below.

RED= Unfiltered raw DSD. Note how the noise slowly rises from 25kHz upward with a peak around 55kHz. Everything past 25kHz is noise inherent to DSD. The level of noise is higher than anything which be could real sonic information (real sonic information tends to exist only at lowish levels past the range of human hearing.) This was frequency analysis taken from a small section Fleetwood Mac - Rumors from the SACD. So much for DSD being a hires format. It's more like CD with more bits and increasing levels noise past 25kHz.

GREEN = DSD filtered per standard practice. This is to keep the high frequency noise down. Note there's hardly any information after 40kHz. This was a frequency analysis taken from a section of Depeche Mode - Strangelove SACD but with the post filter applied (this post filter exists in every SACD player - remember, the DSD information past 25kHz increasingly starts to become noise rather than real sonic information.) So much for DSD being a hires format.

CYAN= PCM 24/192. Looks like we hit the noise floor of the original source by 55kHz. Up to this point, it looks like we have real sonic information. This was a frequency analysis taken from a small section of Madonna - Papa Don't Preach from HDTracks. It would appear that I would have been better off buying the 24/96 version.

DSD as a hires format = BS

Original_Ken · « **Reply #8 on:** December 14, 2013, 02:57:58 AM »

Quote from: purrin on December 14, 2013, 02:54:35 AM

CYAN = PCM 24/192. Looks like we hit the noise floor by 55kHz of the original sample. This was a frequency analysis taken from a small section of Madonna - Papa Don't Preach from HDTracks. It would appear that I would have been better off buying the 24/96 version.

Funny, I was just preparing something about that exact topic. Psychic ...

Paul Stubblebine and Barry Diament wrote this in another Forum (let me know if this sort of partial copy is a problem):

Quote (selected)

Barry has mentioned that he hears a qualitative difference between the 2X rates (88 and 96) and the 4X rates (176 and 192) and I hear it pretty much the way he describes it. As we go up from 16 to 24 bits, and as we go up from the 1X rates to the 2X rates, I hear a number of specific improvements. When we get to the 4X rates done well (and here I agree again with Barry--easier said than done) it's more of a feeling that we have turned a corner and we are almost dealing with a musical experience rather than a facsimile of a musical experience. And I'll confirm that Keith Johnson has said something similar in several conversations.

I'll go further: Those of us who work in digital audio understand the relationships of sample rate to frequency response, and bit depth to dynamic range. Theory says that higher sample rates allow us to record higher frequencies, and in practice that's true. But here's something that the theory doesn't account for: every time we double the sample rate (up to 4X) the bass gets better. Much better. More dimensionality, more texture, more clarity, better decay, lots of things. I'm just trying to make the point that digital audio is more complicated, and more subtle, than the first-level theory that we all learned.

Paul Stubblebine
Paul Stubblebine Mastering, San Francisco
The Tape Project, LLC
serious student of the audio arts

and

Quote (selected)

Hi Paul,

Great to see a post from you here...

"...But here's something that the theory doesn't account for: every time we double the sample rate (up to 4X) the bass gets better. Much better. More dimensionality, more texture, more clarity, better decay, lots of things. I'm just trying to make the point that digital audio is more complicated, and more subtle, than the first-level theory that we all learned..."

Amen!
This is a great point. I've always responded to comments on 4x bandwidth being "higher than humans can hear" by pointing out the relationship between bandwidth and its effect on speed within the audible range. But you raise another important point: the bass does get a whole lot more real sounding at 4x rates (properly done) and this seems to elude the common theory.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

anetode · « **Reply #9 on:** December 14, 2013, 03:11:57 AM »

Those quotes would carry more weight if the respective gentlemen were actually well acquainted with the theory they claim to know better than.

Almost every time I see high def audio discussed on online forums people usually bring out these expert quotes which are based in opinion and unsupported by actual math, rather only what the respective expert feels ought to be true. Sometimes these are supported by utterly misrepresented studies, like when Nugent goes on about jitter audibility.

To be clear: I'm not claiming expertise myself or deriding these people's works as engineers, but when I hear bullshit like "every time we double the sample rate (up to 4X) the bass gets better" I can't help but roll my eyes.

Author Topic: DSD vs CD vs High-Res PCM (Read 14946 times)

Original_Ken

DSD vs CD vs High-Res PCM

Anaxilus

Re: What is DSD?

Marvey

Re: What is DSD?

Original_Ken

Re: What is DSD?

ultrabike

Re: What is DSD?

Original_Ken

Re: What is DSD?

ultrabike

Re: What is DSD?

Marvey

Re: What is DSD?

Original_Ken

Re: What is DSD?

anetode

Re: What is DSD?

(adsbygoogle = window.adsbygoogle || []).push({});

Author Topic: DSD vs CD vs High-Res PCM (Read 14946 times)