Image noise versus camera parameters

I have noticed that there is a huge "boundary layer" beyond which a digital image will have significant noise if blown up by 50% or more. This boundary appears to be defined by ISO, shutter speed, aperture setting, and probably ambient light. I am sure it is also lens and camera dependent. Are there any articles on how to calculate this noise boundary? I use a 1DM2 and a Canon 500mm f4 lens. I appreciate the advice. Thanks.

ps: I am not interested in how to photoshop the solution.I want to learn how to avoid the conditions while composing the image.

One key (often ignored) to minimizing noise in digital capture is to expose for the highlights. Be sure not to clip any highlights but also add exposure if there's headroom such that the histogram is weighted towards the highlights without clipping. This will minimize the shadow areas where noise is most prevelant. There are plenty of well written articles on this subject - I would search for "expose for highlights" in Google or any good website (e.g. Luminous Landscape).

Ernie

William - I'm not quite sure what you're referring to. Are you saying that noise doesn't get progressively worse for you (as exposure gets longer, ISO goes higher, etc), but instead makes a big jump at some point?

Thanks, Ernie. I'll do a web search.

David, no it is progressive as I crank up the ISO and/or increase the speed. I thought there should be a numerical method to determine where really bad noise would occur at given settings (like a three dimensional plot of ISO, Speed, and Aperture). I'm probably imposing my chemistry background in calculating ternary alloys in an area where it does not apply. Ernie may have the best idea. Thanks.

Yeah - Ernie's suggestion is definitely what you should be doing. Even given that, I suppose it would definitely be possible to create a noise plot. I just haven't seen it done. It would be kind of nice to know that the noise at, for instance, ISO 400/0.5s/f3.5 is the same amount of noise as ISO 800/1.3s/f5.6.

Absent an actual plot, the secondary problem you'd have William, would be the definition of "really bad noise". I guess if you knew one setup that gave you unacceptable noise, and had a graph or table that correlated, you'd know where all the limits were.

I wonder if noise is simply a direct correlation with ISO and long exposure. For instance, is the noise at ISO800/2.0s the same as ISO1600/1.0s? I'd guess it's not quite that simple, especially at really high ISOs.

The important thing is to learn the limits and tradeoffs for your gear. If reducing it to plots helps you do that, then the effort will be worthwhile. In the end you need to know what conditions and camera settings will result in noise that's unacceptable to you, then make adjustments accordingly. This might include moving onto a tripod and shooting brief pauses in the action to allow longer exposures at lower ISO's, or in some circumstances fill flash or even full flash illumination might be more appropriate.

It's all very useful info for your decision making, and knowing the parameters of our gear and how to work around difficult situations pretty much runs our professional shooting setups.

You can find some interesting reading on noise HERE.
The link at the bottom of the page also leads you to an informative and well written article HERE

Excellent resources Justin!

Thanks for pointing them out.

Your welcome Hank.
I thought some of the articles on that site were very good indeed.