[shuttertalk]

I suppose an alternative to shooting at higher ISO (with higher noise) would be to shoot at a lower ISO and underexpose, and then adjust later through post-processing. Anyone tried this technique before? Good results?

[adam]

You can see the noise in the shadows when you make it brighter.
Just as long as you don't underexpose too much such that it clips off the end of the histogram.

Many times, I underexpose a little to let me use a faster shutter speed, then correct it later.

[EnglishBob]

Either way you end up with noise... I've tried both, can't say which way works better though.

[Petographer]

I've never under exposed a shot to keep from going to a higher iso. Usually if I have to go to a higher iso it is absolutely necessary.

[Toad]

If you shoot RAW - this may be a good alternative - because you can significantly increae exposure afterwards - however you will still get noise.

[slejhamer]

There is more shadow noise in an underexposed shot that is made brighter in post-processing than there is in a properly exposed shot taken at higher ISO, which in turn has more noise than a shot which is slightly overexposed (but not clipped) and then made darker in post-processing.

Roger Cavanagh did a web article in which he examined the "expose to the right" philosophy: http://www.rogercavanagh.com/helpinfo/28...ight-1.stm

Note that his Page 5 has comments from yours truly.

[Kombisaurus]

hey slej...
here we go again, hey? hehe
just kidding..

I have nothing really to add here (apart from saying that I can see the point in the "expose to the right" theory), but I do have a related question that someone might be able to answer.

How is the "ISO" implemented in the camera?

I would imagine that the lowest ISO represents the true sensitivity of the camera, and that selecting any higher ISO will simply multiply/amplify the signal already received by the sensor. The reason noise is increased is because the noise is also multiplied/amplified.

Is this correct?

And if so, where does this amplification take place? At each pixel site on the sensor? Just as the signal leaves the sensor? During the initial signal processing (ie by the software, but while the signal is still a 12-bit signal)?

Obviously the earlier it is done the better in terms of maximising s/n ratio.

I'm just a bit curious, and I also wouldn't be suprised if it varied a bit from camera to camera (especially the cheaper ones, where I reckon a lot of stuff like ISO is simply software auto-equalisation done right at the end).

Cheers
Adrian

[slejhamer]

hey slej...
here we go again, hey? hehe
just kidding..

I would imagine that the lowest ISO represents the true sensitivity of the camera, and that selecting any higher ISO will simply multiply/amplify the signal already received by the sensor. The reason noise is increased is because the noise is also multiplied/amplified.

Is this correct?

Yes.

And if so, where does this amplification take place?

I haven't found a definitive reference, but my understanding is that it is amplification of the sensor's analog output prior to A/D processing. (Yes, digicam sensors are analog devices ... It's the A/D converters that blow it all to bits. )

[shuttertalk]

You're right regarding the ISO sensitivity and signal amplification. Regarding where it occurs... err.. somewhere within the big metal/platic thing you hold in your hands.

[Don Schaeffer]

It has been my stock in trade (to the extent that I have a stock in trade). Add gamma and contrast. Sometimes you need to blur and re-sharpen or use a noise reduction tool. Sometimes it comes out looking like a painting. When you switch to monochrome it sometimens makes an effective tri-x like print.

It all depends on what the photo is about. We are so pre-occupied with perfect technique that we get bogged down in the safest subject matters. I am glad this question came up because it suggest someone has been playing around outside the box.

--Don