Apache-Talk @lexa.ru 

Inet-Admins @info.east.ru 

Filmscanners @halftone.co.uk 

Security-alerts @yandex-team.ru 

nginx-ru @sysoev.ru 




      :: Filmscanners
Filmscanners mailing list archive (filmscanners@halftone.co.uk)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[filmscanners] RE: Density vs Dynamic range

> > Yes, it is my web site, and I did put that diagram there.  It is
> > straight
> > out of Higgins book "Digital Signal Processing in VLSI".
> >
> That explains a lot.  That diagram (and some of the other things you have
> written) describe the situation with respect to digital signals only.  As
> I've said before, for digital signals the noise is +/- half the step size
> - or one step size.  On that basis your statement that 5000 values are
> required to represent a 5000:1 range is correct, as is the formula you
> quote as the noise is equivalent to one bit.
> But that is only half the story - there is a further analogue noise figure
> that you seem to disregard which, to me, is the far more important figure.

Hi Peter,

Your assessment is not correct.  I have not disregarded the analog noise at
all, in fact, all I have been talking about IS based on the analog noise
level, and nothing to do with digital...it's simply what you NEED in the
digital world to represent the analog signal.

> You can always increase the digital dynamic range (or lower the noise
> figure - same thing) by using a higher resolution AtoD.

If you are already resolving TO the analog noise level, then using a higher
resolution A/D won't increase the dynamic range of the "system".  Lowering
the noise AND using a higher resolution A/D will...

> > But Peter, I am the definitive source ;-)  If you want me to pull out my
> > qualifications, I am more than happy to.  But seriously, please, take my
> > word that what I am describing and talking about I am entirely correct.
> I'm sorry, as an engineer I take nobody's word for anything.  If you want
> to convince me you will have to produce a water-tight explanation.

Not a problem, but just "pretend" what I've said is right, and try to
understand it...

> What you have called "Dynamic range" is really only the range of possible
> values on the digital side.

No, I'm actually talking ANALOG dynamic range, and what it takes in digital
to fully represent that analog dynamic range.

> This may or may not bear any relationship to
> the capabilities of the CCD.

One matches the number of bits in the A/D TO the analog dynamic range.  If
your CCD has a noise level of +-32 electrons, and a saturation of
250,000...that gives it an ANALOG dynamic range of 250,000/64 or 3096...or
36dB...  To represent a dynamic range of 3096, you need 12 bits...(which is
4096, but 2048 won't do it).

> Manufacturers can give us more digital bits,
> but if the bottom 10% are all in the analogue noise region and the top 10%
> represent values beyond the saturation point of the CCD then quoting a
> "dynamic range" based on the digital capabilities of the AtoD are
> nonsense.

Exactly what I've been saying.  Somehow, I think you're reading something
into what I'm saying that I'm not saying, or not reading something into what
I am saying that I am saying...



Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 


Copyright © Lexa Software, 1996-2009.