Archived from groups: rec.photo.digital.slr-systems (
More info?)
In message <42aca0cd$0$41895$ed2619ec@ptn-nntp-reader03.plus.net>,
"Craig Marston" <binaries.newsgroup@craignospammarston.com> wrote:
>> True, but also, using a higher ISO doesn't always mean introducing more
>> noise. If the contrast is low in a scene, and you shoot at ISO 1600 and
>> +2 EC, instead of ISO 400 with 0 EC, the exposure on the sensor is the
>> same, and the noise in the sensor is the same, but the the ISO 1600
>> image will be less posterized, and therefore, more editable, with
>> less-pronounced noise.
>> It isn't the ISO setting that causes noise - it is the exposure level in
>> the sensor, and the intensity of the noise relative to it. High noise
>> and high ISO settings are both symptoms or results of low absolute
>> exposure levels.
>Perhaps I'm misunderstanding what you are saying John, but that makes
>absolutely no sense to me at all!! LOL
That's not surprising; common wisdom suggests that noise comes directly
from the ISO setting.
>If you crank up the ISO sensitivity you are amplifying the output from the
>sensor to a greater degree, which will also amplify the noise [inherently
>present in the sensors*] to a greater degree.
Yes, but you are also amplifying the signal the same amount, so the S/N
ratio is still the same, at the analog level.
>It's about signal to noise
>ratio: the signal which in our case is light needs to be as high as
>possible, and the amplification as low as possible (thus less amplification
>of the noise), to gain the highest signal to noise ratio.
>Am I missing something..?
Yes; the amplifier amplifies both the signal *and* the noise. The
amplification does not determine the ratio; the absolute exposure
(lighting of subject, f-stop, and shutter speed) does. The difference
is, the lower the ISO setting on the camera is, the less RAW, digitized
numbers there are to represent the subject, therefore posterizing it
more.
>*Telescope CCDs are cooled with liquid nitrogen to reduce the amount of
>noise they generate.
Yes, the less noise in the sensor, the higher the S/N ratio, all other
things being equal.
--
<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><