dead pixels?

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

blekenbleu
Posts: 146
Joined: Sat May 10, 2008 5:37 pm
Location: U.S.
Contact:

Re:

Post by blekenbleu »

rjlittlefield wrote:
Tue Jan 28, 2020 10:58 am
Any ideas about why things work so well?
I have not worked in that industry for years, but CMOS sensors were manufactured in fabs converted from obsolete memory chip manufacture.
Unlike in other industries, fabs do not become obsolete from worn out equipment.

Instead, wafer defect rates are reduced to the point that smaller geometries and larger chips become economical.

The dominant defect is reverse-biased photodiode leakage current, which generates so-called fixed pattern noise.

That defect is mitigated in consumer digital camera logic basically by subtracting suitably rescaled dark (shutter closed) pixel values from exposed pixel values.

I do not know whether/which manufacturers ship cameras with actual dead (stuck hot or otherwise) pixels, which would show up in camera raw images,
but suppose that they are quite rare, because many so-called pixel peepers would complain about them on the Internet.

It is probably why reviewers are often not allowed to test raw images from pre-production cameras.

According to a credible blogger, the most prevalent sensor-related camera image defect is photon response non-uniformity (PRNU),
of which most seemingly results from slight gain mismatches among a dozen or so analog-to-digital converters running in parallel to scan pixels into images.
PRNU is becoming diminishingly trivial, buried in photon statistics, impacting Jim's enthusiasm to continue measuring it for new cameras.
https://blog.kasson.com/the-last-word/s ... t-iso-100/
Metaphot, Optiphot 1, 66; AO 10, 120, and EPIStar 2571
https://blekenbleu.github.io/microscope

Post Reply Previous topicNext topic