CD Quality Is Not High-Res Audio

[rbbert]

Just a couple of points that seem to need making repeatedly.

First, if you don't statistically analyze your data both to prove a difference (if present) is significant or not AND to prove a lack of difference is significant, you aren't really "proving" anything with your data (i.e. study) - further remember that in studies of this sort "proving" usually means 95% certainty, not 99 or 100%. That's probably good enough, but still...

Second, a 320k MP3 is closer (in actual data flow rates) to CD than CD is to SACD or 24/192 LPCM.

 

[NorthStar]

Just a couple of points that seem to need making repeatedly.

First, if you don't statistically analyze your data both to prove a difference (if present) is significant or not AND to prove a lack of difference is significant, you aren't really "proving" anything with your data (i.e. study) - further remember that in studies of this sort "proving" usually means 95% certainty, not 99 or 100%. That's probably good enough, but still...

Second, a 320k MP3 is closer (in actual data flow rates) to CD than CD is to SACD or 24/192 LPCM.

Both good points.

 

[Groucho]

Second, a 320k MP3 is closer (in actual data flow rates) to CD than CD is to SACD or 24/192 LPCM.

What is the significance of this?

To my mind, the scale is distinctly nonlinear: CD at 1.4Mbps is comfortably 'good enough'; there's plenty of scope for re-allocating the bits more intelligently for most music, therefore while 320kbps MP3 is just slightly short of perfection, it's probably not enough to notice most of the time (other algorithms e.g. Ogg Vorbis may be better than MP3). But the difference between CD and 24/192 playback is definitely going to be inaudible - what would you bet on your own ability to hear a difference without knowing what you're listening to? (What I bet on your ability to do it? Nothing!) Therefore I don't see any significance in the above point that, apparently, should be repeated over and over.

A question: would you rather have 16/44.1 at 1.4 Mbps, or 24/192 lossily compressed to 1.4 Mbps? Rationally, you might opt for the compressed version (just as you might prefer a photograph stored as jpeg at mega-resolution than a smaller uncompressed version - there probably is more useful information in the compressed version and it could definitely be blown up bigger), but you are committing to trust the compression algorithm that was used and, if later you notice a problem with it, it's too late. If 16/44.1 can be shown to be 'good enough', you might be wiser sticking with it. (Obviously lossless compression doesn't present that dilemma. But what's that you say? Losslessly compressed files sound different from the original? Aaaggh!!!)

 

[rbbert]

I really don't know how significant that point is, part of the reason I put it second. But your discussion isn't really of value unless you think 16/44.1 is as good as it gets, which is what is in dispute.

 

[Johnny Vinyl]

I somewhat agree with Groucho when he says:

while 320kbps MP3 is just slightly short of perfection, it's probably not enough to notice most of the time

I'll be very honest and say that I too cannot always tell the difference. And I know for sure I'd fail miserably at any sort of blind test.

 

[Robin Hood]

Welcome to the Whats Good Enough Forum!

 

[microstrip]

Welcome to the Whats Good Enough Forum!

I prefer the Whats Slightly Short of Perfection - should we carry a poll on WSSP versus WGEF?

 

[Johnny Vinyl]

How about "What's With The Attitude Forum"...seems to be a lot of that going on around here.

 

[DonH50]

It may be worth noting that "what's best" is not often determined by a single number, or in this case pair of numbers. High sampling rates introduce their own problems, like nonlinear settling, higher-speed (and often higher-amplitude) glitches, higher noise, etc. compared to their slower brethren. Higher resolution might not mean much if the noise and linearity does not track, and it usually doesn't. The ideal noise and distortion floor of a 16-bit converter is around 144 dB, not sure I have seen any audio converters hit that yet... There are also the analog buffers going into the ADC and on the DAC output that may dominate the sound (and the measurements). A system with higher-resolution and higher-speed DACs may or may not measure and/or sound better; there are lots of variables beyond the number of bits and how fast they fly by.

For myself, based on comments here, I should definitely get off WBF and find the WGE Forum...

 

[NorthStar]

WBF is only a guide, a reference; not the absolute best. :b ...We can only aspire. ...And inspire. ...And finally expire.

 

[Johnny Vinyl]

It may be worth noting that "what's best" is not often determined by a single number, or in this case pair of numbers. High sampling rates introduce their own problems, like nonlinear settling, higher-speed (and often higher-amplitude) glitches, higher noise, etc. compared to their slower brethren. Higher resolution might not mean much if the noise and linearity does not track, and it usually doesn't. The ideal noise and distortion floor of a 16-bit converter is around 144 dB, not sure I have seen any audio converters hit that yet... There are also the analog buffers going into the ADC and on the DAC output that may dominate the sound (and the measurements). A system with higher-resolution and higher-speed DACs may or may not measure and/or sound better; there are lots of variables beyond the number of bits and how fast they fly by.

For myself, based on comments here, I should definitely get off WBF and find the WGE Forum...

Good post Don!

BTW...if you find the WGE forum lemme know.

 

[microstrip]

Just a couple of points that seem to need making repeatedly.

First, if you don't statistically analyze your data both to prove a difference (if present) is significant or not AND to prove a lack of difference is significant, you aren't really "proving" anything with your data (i.e. study) - further remember that in studies of this sort "proving" usually means 95% certainty, not 99 or 100%. That's probably good enough, but still...

Second, a 320k MP3 is closer (in actual data flow rates) to CD than CD is to SACD or 24/192 LPCM.

Considering that loss-less compression typically achieves 50% perhaps it is more meaningful to normalize to storage rates using loss-less formats than actual data flow rates. Than CD becomes even closer to 320k MP3.

 

[rbbert]

Considering that loss-less compression typically achieves 50% perhaps it is more meaningful to normalize to storage rates using loss-less formats than actual data flow rates. Than CD becomes even closer to 320k MP3.

Along these lines I've noticed that FLAC does not compress 24 bit files to the same extent as 16 bit. depending on program material 50% compression is often possible with 16 bit sources, but 65% is about the best I've seen with 24 bit.

 

[rbbert]

... There are also the analog buffers going into the ADC and on the DAC output that may dominate the sound (and the measurements)...

It's exactly in this area that higher sampling rates and bit depth should make the most difference

 

[NorthStar]

It's exactly in this area that higher sampling rates and bit depth should make the most difference

Sampling rate is more meaningful (audio impacting) than bit depth. ...20-bit is good enough, and anything less than 88.2 kHz is cheap whiskey. :b

 

[microstrip]

(...) and anything less than 88.2 kHz is cheap whiskey. :b

Many people will tell you that cheap whiskey induces headache - please do not tell me that you also consider that CD induces headache?

 

[NorthStar]

Many people will tell you that cheap whiskey induces headache - please do not tell me that you also consider that CD induces headache?

Digital can do that to you; analog not, or less.

 

[Robin Hood]

When I saw the interest of some for the “What’s Good Enough” forum, I thought of the comparisons of the Stereo Review magazine to The Absolute Sound or Stereophile magazines decades ago. I wondered what ever became of Stereo Review and if they had a modern day blog. I traced the history and found that Stereo Review magazine had become the Sound & Vision magazine. I looked online and found the article that started this thread, “CD Quality Is Not High-Res Audio”. http://www.soundandvision.com/node/100295

So now I know where some folks are coming from and know where to send others for more.

 

[Johnny Vinyl]

When I saw the interest of some for the “What’s Good Enough” forum, I thought of the comparisons of the Stereo Review magazine to The Absolute Sound or Stereophile magazines decades ago. I wondered what ever became of Stereo Review and if they had a modern day blog. I traced the history and found that Stereo Review magazine had become the Sound & Vision magazine. I looked online and found the article that started this thread, “CD Quality Is Not High-Res Audio”. http://www.soundandvision.com/node/100295

So now I know where some folks are coming from and know where to send others for more.

Do you make threadcrapping a normal part of your day?

 

[NorthStar]

When I saw the interest of some for the “What’s Good Enough” forum, I thought of the comparisons of the Stereo Review magazine to The Absolute Sound or Stereophile magazines decades ago. I wondered what ever became of Stereo Review and if they had a modern day blog. I traced the history and found that Stereo Review magazine had become the Sound & Vision magazine. I looked online and found the article that started this thread, “CD Quality Is Not High-Res Audio”. http://www.soundandvision.com/node/100295

So now I know where some folks are coming from and know where to send others for more.

Actually, Mark Fleischmann, the author of that short article is a Home Theater guy (small scale). ...He write for few Home Theater mags.

* Sound&Vision is a lower class audio/video mag for the masses; not a serious "audiophile/videophile'' standard mag.
Their product's reviews are the mass market type (from mainly China and Taiwan and Korea and Philippines). ...Onkyo, Pioneer, Sony, Denon, Emotiva, Yamaha, Samsung, Panasonic, ...