Bias value of Canon DSLRs?!

In article <Xns9AB8CBDECC463jpsnokom DeleteThis @130.81.64.196>, John P Sheehy
<jps DeleteThis @no.kom> writes
>Kennedy McEwen <rkm DeleteThis @nospam.demon.co.uk> wrote in
>news:ZpCYTRBbIVTIFwTG@kennedym.demon.co.uk:
>
>>>Most cameras as far as I know get
>>>that
>>>compensation with a dark frame, if desired.
>
>> The only time a "dark frame" is used with the Canon cameras is when long
>> exposure noise reduction is enabled in CF-02.
>
>Hence, his "if desired".
>
Which, in the case of the Canon cameras is not the case for the majority
of the exposure range.

>You're never too smart to slow down and read!
>
Quite, you should try it sometime!
--
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers (replace 'nospam' with 'kennedym' when replying)

In article <Xns9AB8CB6FAA29Djpsnokom.RemoveThis@130.81.64.196>, John P Sheehy
<jps.RemoveThis@no.kom> writes
>Kennedy McEwen <rkm.RemoveThis@nospam.demon.co.uk> wrote in news:C5yeX9AaIVTIFwQ1
>@kennedym.demon.co.uk:
>
>> Since it is analogue it can't be "calculated", whether on the fly or
>> otherwise. What is being offset is the analogue voltage input to the
>> ADC, the digital output is only the consequence of that offset.
>
>The camera still has to decide what that offset will be, to get the analog
>zero signal mean to correspond to the target digitized value.

I don't believe it has to "decide" anything of the sort.

> The 10D was
>inconsistent, the newer cameras are not,

You're sure about that?

> so the newer cameras are probably
>looking at some unexposed pixels, or doing something else like a self-test
>or factory calibration to get the desired results.
>
For many of the reasons given already, a factory calibration would be
likely to achieve anywhere close to the precision necessary - 1 part in
4096. Analogue electronics just isn't that stable.

A self test would be "observable" to the user, especially at long
exposures.

Unexposed pixels might be possible, but again, I doubt this would
achieve the expected consistency. Unexposed pixels are outside the edge
of the frame - a look at the black level of pixels inside the edge of
the frame, especially at long exposures, shows a significantly different
mean black level from the central area.
--
Kennedy
Yes, Socrates himself is particularly missed;
A lovely little thinker, but a bugger when he's pissed.
Python Philosophers (replace 'nospam' with 'kennedym' when replying)