mediasorcerer: The problem with pushing the tone curves is that highlights get blown if some areas are already well exposed, to fix one area often results in another area becoming blown out, and vice versa, you can see this in the chick and bloke pic, image to the right the background sky area is blown out. Still, used with caution it can make a huge difference and help, this is why its best to get the pic as right as possible in the first place, or shoot biased to the left, as far as im aware. Correct me if im wrong.Give the bloke a drink who invented bracketing!!
The point is that those highlights had been blown anyway if you had used the high ISO.
But by using a lower ISO and pushing in processing of the file, they will only be blown in the resulting output, not in the image file. This gives you an option of applying another tone curve to the image file and thus rescuing highlights which would have been lost if you had used high ISO.
coyot3: I tested on my nikon d3100 and the results are at iso 100 1" f4 underexposing 4 ev and later recovering in Darktable(linux) give less noise than iso 3200 1/30 f4 no exposure changes.
I dont know if i understand right the article xD im going to re read ._.
¿Some thoughts on my results?
Coyote, you took two photos at different shutter speeds. That is not how you test ISO invariance.
You need to take some raw photos with the same shutter speed, same f-stop and same scene light, but different ISO. The low-ISO photos will be darker, so you brighten them to match the brightness of the high-ISO photos and look for differences.
The test should be done on a subject which has deep shadows in the high-ISO photo since this is where you will see the largest differences if your camera isn't ISO-invariant.
Simon Simpson: Quoting from the first paragraph in the article:
“the randomness of the light” – with respect there is no such thing; there are, however, quantisation errors of camera sensors which lead to variations (and thus noise). Quantisation is the process of turning light into electronic signals. I’m not sure the first sentence in the article makes complete sense ?
“it's much easier to think of noise the way you think of background 'hum' from a Hi-Fi amplifier.” – hum is not a random noise derived from the electronics; hum is an unwanted audible effect derived from an alternating current household electrical supply to electronic equipment, normally audible at a frequency between 50–60Hz. Noise is randomly generated by the electronics and is typically described as ‘hiss’.
Just thought I’d clarify that.
Simon, randomness of light certainly exists, and it should not be confused with quantisation errors.
If you had a sensor and processing chain which formed a perfect photon counter, there would be no quantisation errors, but you would still have noise from photons arriving randomly instead of arriving at fixed intervals.
People often think of light as a constant, homegeneous flow, but it isn't. Light consists of single photons, and they arrive at completely random intervals. If you have a constant light source, and the photons arrive at a sensor pixel an average interval of for example 1000 microseconds, it is perfectly possible that no photons arrive over a given period of 3000 microseconds, or that 5 photons will arrive over a given period of 1000 microseconds.
Impulses: Nice additions, both of them, it'd be nice if the mobile UI got some love too tho. I usually detest mobile sites but your actually works reasonably well and I've been using it for a while, particularly for the boards... BUT, there's two things that would make a world of difference: mobile sorely needs an edit button (forums and comments) AND the mobile forum really need a shortcut to skip to the last page of a thread, like seriously.
No. It needs a shortcut to skip to the first unread message in a thread. There is a large difference.