In article , Andrew Rose at Pristine
Audio wrote:

"Jim Lesurf" wrote in message



If they "know what they are talking about" I am puzzled that the page
whose URL you gave contains the uncorrected errors/misunderstandings
which I found there.

It's a discussion in an online forum. I don't censor opinions or points.
Why therefore should you be puzzled?

For the reason I gave. The page contains some fairly clear and basic errors
and misunderstandings. If the contributors "know what they are talking
about" I would expect that either:

1) The errors/misunderstandings should not have appeared in the first
place.

or

2) They would have been promptly followed by corrections/explanations that
clear up the errors/misunderstandings.

I was not referring to a lack of 'censorship' but to the evidence that at
least some of those involved show a lack of "knowing what they are talking
about" and that this does not seem to have been promptly corrected by
others who "know what they are talking about".

I accept you aren't to blame for what contributors may write. Nor should
you be 'censoring' as that implies control of opinion and preference as
distinct from ensuring factual accuracy and clarity. The problem is not
with 'opinions' stated, but with some of the statements being incorrect
according to information theory, and the relevant evidence/practice.

This does not mean I am saying that everything on the page is wrong, or
that some contributors don't have any relevant knowledge. I'm just
expressing surprise as a result of the fairly basic nature of the errors/
misunderstandings not fitting with your comment about the knowledgeability
of the contributors.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

In article , Stewart
Pinkerton
wrote:
On Wed, 16 Feb 2005 17:45:51 +0100, Andrew Rose at Pristine Audio
wrote:

In message , Arny Krueger
writes


*Any* properly dithered digital signal has zero nonlinear distortion.
That is zero, as in zero followed by a decimal point followed by an
unlimited number of digits for the fractional part, all zero. IOW the
distortion is infintesimal, and not finite.

Is this true right down to, say, 8-bits? If I produce a sine wave at
-60dB in a 32-bit environment and dither down to 8-bits does this still
stand?

Yes, it is true. Indeed, it's true right down to one bit, as should be
obvious from the ubiquity of one-bit converters! You can't remove
distortion with noise shaping................

The only problem you'll have with your -60dB signal in an 8-bit
environment is that it will be some 12dB below the wideband noise floor.
You can however still extract it with a narrow-band analyser - or a pair
of ears! :-)

To expand on the above:

If we reconvert/requantise from 16-bit to, say, 8-bit, then we also have to
use a higher sample rate for the resampled/converted stream in order to
keep all the original information. This also means that although we now
have a *wideband* noise level appropriate for 8-bit, we can choose to
arrange for the bulk of this noise to be *outside* the original bandwidth
of the 16-bit sample sequence. Hence in the audio band, the result may not
show as being noisier than the original 16-bit version. This is one of
the areas where 'noise shaping' is relevant.


OTOH If you simply requantise down to 8-bit but do not use a higher
sample rate for the results, then the noise level will rise if you
wish to dither correctly. Otherwise you will simply introduce quantisation
distortions which correct dithering would prevent. This is inevitable
as the 8-bit stream - if at the same sample rate as the 16-bit one - cannot
convey the same information rate, hence must have more noise and/or
more distortion.


Beyond the above generalisations, the results will depends entirely
on the details of the process.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html