Enhanced 20 Bits Surround sound? Options

[JEN]

On the back of 1 of my CDs, it says "enhanced with brilliant effects mastered digitally in 20 bits Surround sound"

What does this mean?

[Sachankara]

It means that you as a consumer should believe that it's superior to a normal CD. Pseudo scientific "hidden audio" that won't be noticable even if you have a compatible player...

This post has been edited by Sachankara: Mar 8 2003, 21:15

[CiTay]

Pseudo scientific "hidden audio" that won't be noticable even if you have a compatible player...

I don't think this has anything to do with the final product; they are talking about the mastering stage. It depends on many more factors if the end result sounds good or not (which is of course 16 bit).

[JEN]

So... if I encode it to a lossy format, will it stay as 20 bits or will it become 16 bits?

How about lossless formats?

[CiTay]

So... if I encode it to a lossy format, will it stay as 20 bits or will it become 16 bits?

How about lossless formats?

Read again what i wrote above

What you have uses 16 bits, not more.

[Guest_SK1_*]

I've never heard of 20 bit for any stage... 24, 32, 64 yeah, but not 20...

[JEN]

Oh...

[CiTay]

I've never heard of 20 bit for any stage... 24, 32, 64 yeah, but not 20...

This is not so uncommon for audio hardware. For instance, my soundcard has an 18 bit ADC and a 20 bit DAC.

[JEN]

OK

This post has been edited by jenny: Mar 8 2003, 21:36

[sven_Bent]

i have a freind that once beleived that many cd's was with 20 bits but only VERY good cd palyers coud lread them because they where more difficult to read then the first 16bits

also he called himsel audiophile. and would not belive when i told that the cd foramt only supports 16/44/2 format. and nothing else.
what can i say...audiophiles :-)

[Bedeox]

@jenny: It is either marketing gibberish or HDCD.

@sven_Bent: Poor him. Did he hear the difference?

<edit>
Stupid me, I didn't mean SACD but HDCD.
</edit>

This post has been edited by Bedeox: Mar 9 2003, 11:04

[tigre]

Applying dither to a cd mastered at 20bit+ in a final step increases the possible dynamic range (not S/N ratio) of a CD compared to plain undithered 16bit audio. If it's noise shaped dither and you look at the lowest volume where a tone (frequency arround lowest point of ATH curve) can still be noticed besides dither noise, you could say the dynamic range is the same as for 20bit undithered.

Or maybe it's HDCD?

This post has been edited by tigre: Mar 8 2003, 23:04

[tigre]

... you could say the dynamic range is the same as for 20bit undithered.

This was what I was *thinking* - but it made me curious, so I did a test (maybe a bit off-topic):

I created a 0dB 1000Hz tone in CEP (32bit resolution) and amplified it by -96dB, -99dB, -102dB ... and dithered the result to 16bit (44.1kHz noise shaped 0.7 bit dither). Afterwards I amplified the result by 60dB. The lowest 1000Hz tone I could hear through dither noise had an amplitude of 0.0207 before dithering to 16bit (max. amplitude 32768). If this would be marketing we could say:

"Enhanced with brilliant effects dithered digitally in 20.6 bits superior 124 dB dynamic range sound"

-edited: some clarifications in description

This post has been edited by tigre: Mar 9 2003, 01:03

[Pio2001]

At a frequency of 1/N the sample rate, you've got N/2 samples to describe how high the curve grows (or how deep it falls). Averaging all samples, it is possible to get a result N/2 times smaller than the elementary step.

So 22 kHz can't have anymore dynamic than 16 bits (96 db)
11 kHz 17 bits (102 db)
5.5 kHz 18 bits (108 db)
2.8 kHz 19 bits (114 db)
1.4 kHz 20 bits (120 db)
789 Hz 21 bits (126 db)
etc

[tigre]

At a frequency of 1/N the sample rate, you've got N/2 samples to describe how high the curve grows (or how deep it falls). Averaging all samples, it is possible to get a result N/2 times smaller than the elementary step.

So 22 kHz can't have anymore dynamic than 16 bits (96 db)
11 kHz 17 bits (102 db)
5.5 kHz 18 bits (108 db)
2.8 kHz 19 bits (114 db)
1.4 kHz 20 bits (120 db)
789 Hz 21 bits (126 db)
etc

I like the similarity between these figures ...
1.4 kHz 20 bits (120 dB)
789 Hz 21 bits (126 dB)

... and my test result:
1.0 kHz 20.6 bits (124 dB) B)

[Frank Klemm]

At a frequency of 1/N the sample rate, you've got N/2 samples to describe how high the curve grows (or how deep it falls). Averaging all samples, it is possible to get a result N/2 times smaller than the elementary step.

So 22 kHz can't have anymore dynamic than 16 bits (96 db)
11 kHz 17 bits (102 db)
5.5 kHz 18 bits (108 db)
2.8 kHz 19 bits (114 db)
1.4 kHz 20 bits (120 db)
789 Hz 21 bits (126 db)
etc

5.5 kHz -- 17 bit (104 dB)
1.4 kHz -- 18 bit (110 dB)
350 Hz -- 19 bit (116 dB)
88 Hz -- 20 bit (122 dB)

You spread _power_ not voltage, and the dynamic of a n bit system is 20 * log (1.5 * 2^n),
because max PP is 2^n, max. RMS is 2^n/sqrt(8), noise is Int from -0.5 to +0.5 of x² dx = 1/sqrt12).
Simple math.

[gdougherty]

At a frequency of 1/N the sample rate, you've got N/2 samples to describe how high the curve grows (or how deep it falls). Averaging all samples, it is possible to get a result N/2 times smaller than the elementary step.

So 22 kHz can't have anymore dynamic than 16 bits (96 db)
11 kHz 17 bits (102 db)
5.5 kHz 18 bits (108 db)
2.8 kHz 19 bits (114 db)
1.4 kHz 20 bits (120 db)
789 Hz 21 bits (126 db)
etc

5.5 kHz -- 17 bit (104 dB)
1.4 kHz -- 18 bit (110 dB)
350 Hz -- 19 bit (116 dB)
88 Hz -- 20 bit (122 dB)

You spread _power_ not voltage, and the dynamic of a n bit system is 20 * log (1.5 * 2^n),
because max PP is 2^n, max. RMS is 2^n/sqrt(8), noise is Int from -0.5 to +0.5 of x² dx = 1/sqrt12).
Simple math.

In the voice of Cronk, from The Emperor's New Groove

"Simple math..... Right...."

Perhaps it's my lack of undestanding of the cryptic acronyms, but I'll leave the simple math to the codec developers and just listen to the music. BTW, Thanks Buschmann and Klemm for all the hard work on Musepack. You guys are the MF Bomb.

[gdougherty]

On the back of 1 of my CDs, it says "enhanced with brilliant effects mastered digitally in 20 bits Surround sound"

What does this mean?

It can mean one of several things.

1) The audio was recorded at 20-bit or higher, or even in analog though that's a really painful way to do things compared to the ease of digital recording and editing. From there it was mastered using equipment that operated at 20-bits then dithered down to 16-bit for the final CD.

2) Same start as the rest except it was encoded with HDCD, which I'm led to believe is very similar to doing the previous option, except it adds a cool hype phrase to the product and attracts "high-end" listeners who demand that their CD's be HDCD when they aren't available on DVD-Audio or SACD.

3) It's a special 20-bit CD not playable in normal CD players or any other CD player I'm aware of. If you can listen to your CD, it's not this option.

G

[KikeG]

Well, HDCD capable players actually decode HDCD discs to 20-bit data. However, that doesn't mean they will sound better.

[budgie]

KikeG:

You're surely joking... It sounds a way better, no need to ABX either...

[KikeG]

You're surely joking... It sounds a way better, no need to ABX either...

Well, I doubt it, mainly for two reasons:

- I don't know of any recording that fully exploits the dynamic range of 16-bit cd format (94 dB in practice).
- I doubt you have a listening facility with less than 20 dB of ambient noise (30 dB is usual at quiet rooms) and at the same time you listen music at volumes over 110 dB peak level, only way where maybe you could appreciate full dynamic range of 16-bit cd format.

Perceived differences are either due to better mastering of the HDCD record, or just placebo. Same as with SACD/DVD-A.

[budgie]

Well, I have these HDCDs (amongst others, more I can't remember for the moment, but there's a lot more...): Bryan Ferry "As Time Goes By"; King Crimson "In The Wake Of Poseidon"; King Crimson "Lizard"; King Crimson "Islands"; Niacin "Deep"; "Big Band Basie"; Yes "Magnification"... Some weeks ago, just for the fun, I connected one of my CD players via AUX input and simultaneously into CD input of my amp using digital output of the CD player via X-DAC from Musical Fidelity, which has HDCD decoder built in. The difference in sound was immediate and evident, there's really no need for making even blind tests...

This post has been edited by budgie: Mar 11 2003, 11:08

[KikeG]

The mismatch in output level between the two devices can have easily fooled you. This is quite usual in non-controlled listening tests.

This post has been edited by KikeG: Mar 11 2003, 11:14

[budgie]

KikeG, you make me cry... or am I to laugh? I am experienced enough to distinguish between output levels and better resolution output But it doesn't matter, anyway...
No more posting in this thread from me.

[KikeG]

I am experienced enough to distinguish between output levels and better resolution output  But it doesn't matter, anyway...

You can laugh all what you want, that won't change the fact of that it's not an unimportant issue. A level mismatch of 1 dB is easily noticeable as one of the devices sounding "fuller", there are lots of evicence for this.

The fact is, for a listening test to be reliable, levels must be matched within 0.1 dB of difference (which has been proved to be inaudible), which, of course, is only possible with the help of measurement equipment (any sound card is fine, also a voltmeter plus a continuous test signal can do).

And you can reply as many times as you want, in fact I like discussing such things.