Noise ... exposure time vs ISO

[Peter McLennan]

[font color=\'#000000\']Yes[/font]

[BJL]

[font color=\'#000000\']Thanks Jonathan.

For those who care about the technical details, item 2 is because using the lower ISO and then pushiing digitally involves amplifying less in the analog mode, and then more later in digital domain. Thus, any noise that gets introduced in between these two stages (such as in the A/D conversion itself) gets amplified more.[/font]

[BJL]

[font color=\'#000000\']How common is in camera dark frame subtraction on dslr's?[/font]

[font color=\'#000000\']Fairly common I think; some I know of off-hand are both the Olympus E-1 and the coming cheaper E-300, and the 20D has it too I believe, in place of a different ("no waiting") approach taken in the 10D.[/font]

[Wayne Fox]

[font color=\'#000000\']Just curious if anyone out there has done any study on which introduces more noise ... longer exposure or higher ISO.  In other words, if presented with a choice of a 2 second exposure at ISO 100, or a .25 second exposure at ISO 800, would the longer time yield less noise because of the lower ISO?

This is assuming a good low noise camera to start with like a 1Ds.[/font]

[BJL]

[font color=\'#000000\']I agree that the longer exposure should give better S/N ratio. Can I change the example to 2 seconds at ISO 100 and 1/2 second at ISO 400 for simpler math?

The important point is that the aperture ratio is being held constant due to the adjustment in ISO setting, so light is gathered at the same rate.

The longer exposure will gather four as much light and so give four times the signal. This will produce twice as much photon shot noise (it is a square root law). The dark current noise will also be four times as much, but the spatial variation of it will be only twice as much (anothe square root law).

Thus overall, the noise should be less than four times as much, and so the signal to noise ratio should be better for the longer, lower ISO option. I am ignoring pre-amplifier noise, but that should be similar in either case, and so favors the longer exposure even more.[/font]

[Jonathan Wienke]

[font color=\'#000000\']Some general principles of digital exposure:

1. The ideal exposure level that will capture the most dynamic range with the least noise and other undesirable sensor-related artifacts at a given ISO is that which causes the RAW data to closely approach, but not reach the clipping point for non-specular highlights. AFAIK there are no exceptions to this principle.

2. It is always preferable to increase ISO to achieve ideal exposure than to underexpose at a lower ISO setting and push the exposure in post-processing.

3. If increasing exposure time is an acceptable option (it will not cause an unacceptable degree of motion blur), it is better to decrease the ISO and increase exposure time to achieve the ideal exposure level. The only exception to this principle is when the lower ISO setting captures less dynamic range and that would cause clipping that would not otherwise be encountered. (An example would be ISO 50 on the 1Ds.)[/font]

[Scott_H]

[font color=\'#000000\']How common is in camera dark frame subtraction on dslr's?  The E-20 has it and it makes a huge difference, but it has a pretty small sensor and is noisy to start with.  

I like doing long exposure, and I guess I am wondering about the capabilities of the dslr's on the market of doing 20-30s exposures that are noise free.[/font]

[61Dynamic]

[font color=\'#000000\']

How common is in camera dark frame subtraction on dslr's?

Quite common. The 300D is the exception that I know of. It has what Canon calls on-the-fly DFS, but it does not work worth a dang. Exposures have to be kept below 5-10 min. otherwise hot-pixel removeal becomes a chore.

You could do DFS after the fact in PS however...[/font]