5D MarkII and ISO

I would like to politely inquire where people heard about the notion of some iso being real, and some being "done in software". I am just asking, because I never heard of it, and my notion of how ISO sensitivity works for digital suggests that the amount of gain from the sensor is selectable by say a voltage level, which can be any value between the low and high values. Therefore the ISO values like 100, 200, 400, 800, 1600 are just arbitrary as far as the camera is concerned, they are values that are meaningful to us as humans. The gain could be set to something which equals ISO 50, or ISO 25600, whatever the sensor gain can be set to.

I'm certain someone else can explain it better.

Here's a thread from a year ago or so that has some links:

https://photography-on-the.net/forum/showthre​ad.php?t=657074

Canon doesn't come right out and spell out its "secrets" but a lot of evidence point to the use of "software amplification" for intermediate ISOs rather than direct electronic amplification/gain as is used in the 100-200-400-800-1600 ISO. This is also true about ISOs labeled "H" and "L" -- Canon even gives warnings that these ISOs may have less-than-stellar performance. They are pushing the bounds of exposure. In essence, the results of the tests indicate that the "+1/3" steps are boosting the lower (underexposed) ISO gain and the "+2/3" steps are pulling back the higher (overexposed) ISO gain, resulting in lower noise but possible highight clipping.

ISO 100 is considered "optimal" because it is the lowest "native" (i.e. electrical) ISO and because it is at that level providing at a good exposure the maximum amount of light that the sensor is designed to handle -- in other words, a setting of ISO 50 in reality will overexpose, tend to lose some highlight data before the pull back to a "good" exposure.

I'm typing all this on the fly and don't have a bunch of references on hand, but that's the gist of the research.

It should be noted that some tests of the 1-series cameras indicate that they have more electronic amplification/gain applied at intermediate ISOs, but I don't have any details.

tonylong wrote in post #10234597
...
It should be noted that some tests of the 1-series cameras indicate that they have more electronic amplification/gain applied at intermediate ISOs, but I don't have any details.

I've heard this often quoted on POTN as well, but the only data I could find also shows that the 1 series cameras also interpolate the 1/3 stop ISO values. If you look at this link and scroll down to the noise samples it clearly shows that ISO 400 is better than ISO 320, ISO 800 better than ISO 640 and so on.
http://www.northlight-images.co.uk …ras/canon_1ds3_​noise.html

Full stops or partial stops it doesn't matter, if properly exposed noise is virtually non-existent on the 5D2 up to ISO 3200.

Tony, educational and well said. Thank you.

2nd weird question - I can't seem to get my 1D III to even expose at ISO 100... lol. This means I'm gonna have to read the manual, doesn't it? It goes down to 200 (which is fine, unless I've got too much light and need to shoot wider aperture with fill flash) and that's all. I'm thinking it's buried in a custom function that I haven't taken the time to find yet as it's not a big problem...

collierportraits wrote in post #10236274
2nd weird question - I can't seem to get my 1D III to even expose at ISO 100... lol.. ..

Turn off highlight tone priority.