
CCD Noise and Hot Pixels
7 months ago
OK, so its not the next inde film... but for a couple people it might be interesting stuff. This is a movie of CCD noise over time. I noticed that some of my time lapse work seemed to get more noisy over the hours. So I took about 500 shots in total darkness over a 15 minute period. The raw shots were, as you would expect, black.
I amped up intensity by about 30x so as to bring out the bottom 3-6% intensity and make the noise visible. Then I used edge detection to turn single lit up pixels into small dots more easily seen.
Its sort of hard to see, but over the run time of the clip you can see the noise increases over time. You can also see about a half dozen hot or stuck pixels. This, I plan, to use as a means of incorporating hot / stuck pixel mapping and adjustment to SBG. So worth the time.
As a practical matter, this underscores the need for camera manufacturers to consider heat dissipation in their devices as a more than superficial matter. From the graph of noise over time, which roughly corresponds to the heat up of the camera, we see a 400% increase in noise! That's 1 or 2 stops! This also means that cooling the camera would drastically reduce noise. This, of course, has been known by astronomers for decades.
Anyway, hope this is interesting to someone other than me.
I amped up intensity by about 30x so as to bring out the bottom 3-6% intensity and make the noise visible. Then I used edge detection to turn single lit up pixels into small dots more easily seen.
Its sort of hard to see, but over the run time of the clip you can see the noise increases over time. You can also see about a half dozen hot or stuck pixels. This, I plan, to use as a means of incorporating hot / stuck pixel mapping and adjustment to SBG. So worth the time.
As a practical matter, this underscores the need for camera manufacturers to consider heat dissipation in their devices as a more than superficial matter. From the graph of noise over time, which roughly corresponds to the heat up of the camera, we see a 400% increase in noise! That's 1 or 2 stops! This also means that cooling the camera would drastically reduce noise. This, of course, has been known by astronomers for decades.
Anyway, hope this is interesting to someone other than me.
- Vimeo: About / Blog / Roadmap / Developers / Community Guidelines / Forums / Toys / Help! / Site Map
- Legal: ©2008 Vimeo, LLC / Terms & Conditions / Privacy Statement




.
59
0
0
0
Previous Week