SACD/DVD-A vs. Oversampling

[WmAx]

Now is it really all that? Professional recording/mastering engineers seem to be split on the issue, at least from what I've read over the years. Most still think a well mastered CD that truly utilizes the full 16 dynamic range can sound just as good. The higher bit depth is great, but the higher sampling frequencies (more samples per second) doesn't necessarily buy that much.

It's only fair to point out that those that don't think 44.1/16 is adequate for playback, have not substantiated their cause for concerns in a proper experiment or based there reasons on existing credible research that concludes such.

The increased bit depth gives you finer grained samples. With CD 16 bit word length, the values for each sample (44.1 K per second) range between -32767 and +32,768 [LPCM is signed and the waveform has both positive and negative components]. With 24 bit, you get a range of -8,388,607 to 8,388,608 - so the accuracy of the amplitude of each sample is better.

The increased bit depth is ultimately relavent so far as relative max. signal level vs. min. signal level. The additional values are used to extend the realtive amplitude capability, not to increase the samples in a specified range of amplitude. In a practical sense, each bit is distributed as approx. a 6 db relative increment in both 16 bit and 24 bit current systems. Thus, CD format: 16 bits = 96dB range, where as DVD-A format: 24bits = 144dB range.

-Chris

 

[Buckle-meister]

The increased bit depth is ultimately relavent so far as relative max. signal level vs. min. signal level. The additional values are used to extend the realtive amplitude capability, not to increase the samples in a specified range of amplitude. In a practical sense, each bit is distributed as approx. a 6 db relative increment in both 16 bit and 24 bit current systems. Thus, CD format: 16 bits = 96dB range, where as DVD-A format: 24bits = 144dB range.

Ah, ha! But I read another post that said that in the 80's when folk didn't record so loudly onto CD's, a well recorded piece of music could have a dynamic range of around 25dB.

Why then isn't CD (or vinyl for that matter) more than adequate with amost four times the dynamic range 'required'? And isn't DVD-A therefore overkill?

 

[MDS]

I agree WmAx. I'm one that does think 16/44.1 is adequate, although i've seen my share of overly compressed waveforms with less than 6dB crest factor. Certainly those are not living up to the potential of CD.

 

[sploo]

(sorry Sploo, but I need to borrow the thread again )

I'll fight you for it

can people even hear 1 part in 32767? Sounds dubious to me.

Yeah, you do wonder how much is really audible, and how much is willy-waving (I've got more bits than you!)

Way, way back I remember playing around with sampling on a Commodore Amiga (remember those? ). That had an 8bit sound system, and sample rates were usually around the 8kHz mark.

As MDS pointed out, you can reduce the aliasing with higher sample rates, and of course, more accurate samples (more bits) serve to further reduce the 'jaggies' (OK, so I'm using an old computer graphics aliasing term, but you get my point).

I seem to remember reading an article, probably linked/recommended via this forum, that showed results of listening tests with various sample accuracies - and I think once samples started to get towards 12-14bits people were struggling to reliably hear the difference with originals.

From playing about with computer sampling and playback for over 15 years I'd concur. But maybe too many years of loud music has taken its toll .

EDIT: Again, taking so long to type a reply I miss posts . It'd never occurred to me that the extra bits were used to increase dynamic range, rather than increasing the accurancy within the same range. Still, that's why we're here - to learn!

 

[WmAx]

Ah, ha! But I read another post that said that in the 80's when folk didn't record so loudly onto CD's, a well recorded piece of music could have a dynamic range of around 25dB.

Even current hyper compressed recordings willl have greater than 25 dB actual range. The figure you are referring to is the difference between the average RMS power of the signal vs. the maximum peak value available. Today, only 6-10 dB of peak range are left on many hyper compressed recordings. But, if you analyze the signals, you can still find 35 to 40 dB total range used in even the hyper compressed recordings(not counting the fade in/fade out of the track start/end positions).

Why then isn't CD (or vinyl for that matter) more than adequate with amost four times the dynamic range 'required'? And isn't DVD-A therefore overkill?

You are referring to the range as if it were linear(4x). It is not a linear scale. For example, 96 db is almost 13x the amplitude of 25dB. Remember, every 6db is 2x the initial amplitude. 46dB is twice the amplitude of 40db, for example.

-Chris

 

[mtrycrafts]

And isn't DVD-A therefore overkill?

Well, unfortunately CDs are not in multi channel format
But, yes, SACD and DVD-A is overkill when sampled beyond 48kHz and 18bits as J. stewart of Meridian would say above audibility after that point

 

[Buckle-meister]

Today, only 6-10 dB of peak range are left on many hyper compressed recordings. But, if you analyze the signals, you can still find 35 to 40 dB total range used in even the hyper compressed recordings(not counting the fade in/fade out of the track start/end positions).

Two CD's that I have are Inxs' greatest hits, which is one of my most loudly recorded CD's, and Dire Straits' Brothers in arms, which is easily the most quietly recorded CD that I own.

I have read on another post that Brothers in arms is a well mastered/recorded album; old though it may be.

Does the above quote necessarily mean that, by virtue of its loudness, Inxs' greatest hits has been poorly transferred to CD?

 

[tbewick]

I think that I was wrong in my earlier comment about the criticisms of audio CD being rubbish. There is an excellent article about digital audio on the Rane Audio web site:

http://www.rane.com/note137.html

I'd also recommend a book called 'An Introduction to Digital Audio', written by John Watkinson.

 

[Buckle-meister]

I think that I was wrong in my earlier comment about the criticisms of audio CD being rubbish. There is an excellent article about digital audio on the Rane Audio web site:

http://www.rane.com/note137.html

I'd also recommend a book called 'An Introduction to Digital Audio', written by John Watkinson.

Thanks mate. I visited the main library the other day and could only find a couple of related books on 'servicing CD players' etc. One book did have a nice initial chapter talking about most of the stuff Ive learned from posings here, some with url links referenced within, but mostly they did not cover what I was looking for. I'll check out your reference for sure.

Many thanks.

Kind regards.

 

[dponeill]

I think we may be confusing oversampling with upsampling?

Dennis

 

[WmAx]

Does the above quote necessarily mean that, by virtue of its loudness, Inxs' greatest hits has been poorly transferred to CD?

This depends on the relative loudness. If the Inxs has any sort of thing that should have considerable dynamics, such as drums, acoustic intruments or even voices, then if the average loudness level of the CD is -10dBFs[for example], then it is insufficient, and must result in degraded audio fidelity.
-Chris

 

[WmAx]

Thanks mate. I visited the main library the other day and could only find a couple of related books on 'servicing CD players' etc. One book did have a nice initial chapter talking about most of the stuff Ive learned from posings here, some with url links referenced within, but mostly they did not cover what I was looking for. I'll check out your reference for sure.

Many thanks.

Kind regards.

Are you looking for an understandable overview of digital audio sampling theory? If so, here:

http://www.lavryengineering.com/documents/Sampling_Theory.pdf

-Chris

 

[eirepaul]

Wow, there has been a lot of technical discussion going back and forth on this thread. I won't contribute to that, but I will say, in response to Buckle-meister's original question, that I have several recent DVD-A/SACD versions of 1970s/80s recordings that I also have had in CD format for several years.

It certainly seems to me that the multi-channel formats sound more dynamic and cleaner. Now, whether it is the re-mix job by the engineer, the fact that the music is opened up to details so much more with multi-channel, the DVD-A/SACD technology, additional mastering or probably all of the above, I am not sure. The bottom line is they sound clearer/cleaner (resolution?) to me and I hear a lot more in the recordings. This is even true with stereo SACDs (Peter Gabriel - So is a good example I have).

Not every surround mix is great, I'm sure; but most of them that I have heard have added a lot to the enjoyment of the music.

 

[Buckle-meister]

It's your own fault WmAx!

Are you looking for an understandable overview of digital audio sampling theory? If so, here:

http://www.lavryengineering.com/documents/Sampling_Theory.pdf

Sorry for the delay responding WmAx; I had thought this thread was dead and hadn't been checking up on it. After reading your link (nice ) you will be delighted to know that I now have additional questions (collective groan from the forum!)

1. '[Nyquist] stated that the difference between the original wave and the original waveform is high frequency, thus the key to making them the same (to remove the difference) is filtering out high frequencies (above Nyquist). Any filtering (removal) of the high frequency energy would remove the difference, thus yield the original.' Is this saying that digital recording results in 'more information than what we want', but that the 'extraneous information' is removed upon playback? i.e. mathmatically analagous to saying that instead of 2+2=4; 4=6-2

2. What's 'anti-imaging'?

3. 'The compromise between speed and accuracy is a permanent engineering and scientific reality.' Ha! In essence then, CD's are the most accurate medium (compared to SACD/DVD-A) that we currently have due to slower processing speed? Further; by using only the bare minimum (Nyquist) sampling rate rather than those much higher of SACD/DVD-A, we currently have the best digital recording we can possibly get, ever, and have done since 1987!?

Now, whether it is the re-mix job by the engineer, the fact that the music is opened up to details so much more with multi-channel, the DVD-A/SACD technology, additional mastering or probably all of the above, I am not sure.

...or that you are simply imagining it

No offence; This place turns you into cynic (sigh)

Regards

 

[WmAx]

Is this saying that digital recording results in 'more information than what we want', but that the 'extraneous information' is removed upon playback?

The only context in which there is more information than is desired, is where such exceeds the Nyquist limitation(s).

2. What's 'anti-imaging'?

Essentially the same thing as anti-alias, but not exactly. Usually, anti-alias is used in the same capacity as, or in the place of anti-imaging[the low pass filtering stage after the DAC].

3. ' Ha! In essence then, CD's are the most accurate medium (compared to SACD/DVD-A) that we currently have due to slower processing speed? Further; by using only the bare minimum (Nyquist) sampling rate rather than those much higher of SACD/DVD-A, we currently have the best digital recording we can possibly get, ever, and have done since 1987!?

Depends on the context of accuracy. If you have a single clock, and you compare the timing[jitter] error of frequency x and x*2, and consider accuracy as a degree of percent of error at the maximum reproducable frequency of each, then the lower frequency rate will have lower percentage of error and higher frequency will have a higher percentage error. But if you consider accuracy, for example, as the ability to reproduce an impulse response, then the higher frequency example will be superior.

As far as SACD or DVD-A vs. CD: No credible research has demonstrated a need for anything greater than CD's abilities, for purpose of playback of musical content.

-Chris

 

[Buckle-meister]

Thanks mate; much appreciated.

 

[DR_AUDIO]

The Sound

I like the multichannel format of DVD audio and SACD. There are good and bad recordings of each and the same of CD. There are some wonderful sounding CDs and bad sounding recordings using the 24 bit 192Khz formats. In the end it is the mastering and engineering that is most likely important, however the recent HYBRID SACD's usually show a distinct advatage sonically used in 2 channel SACD format over their CD counterpart. Go figure.

 

[WmAx]

however the recent HYBRID SACD's usually show a distinct advatage sonically used in 2 channel SACD format over their CD counterpart. Go figure.

It may be that the CD layer is the same as the conventional CD version on many SACD hybrids, where as the SACD layer is a different mastered version. I know that upon inquisition of a specific title[Tierney Sutton, Dancing In The Dark], Michael Bishop of Telarc admitted to me that the CD layer on the SACD of the title of mention, was the same mastered version as the CD release, where the SACD layer was a different mastered version[with less dynamic compression] as compared to the CD layer.

-Chris