Scott reminded me about this important concept. Basically, you overexpose on a digital camera so that you lift the darkest regions out of the noise and then in digital processing, you put it all back again. It is kind of like the way Dolby C noise reduction worked in the old analog days (they would increase the quiet souds artificially while recording and then the cassette recording would reduce it, driving the noise down too). This is sadly something not yet built into digital cameras so for most cameras, you have to manually overexpose them and then pull them down in Lightroom or what every you use.
The net is that if your camera supports it one easy way is to just look for the blinking showing things that are washed out, keep exposure compensating until you get there and also you can on some cameras exposure bracket, so you'd want something like taking three shots each time at +2/3EV higher than when you get blinkies.
You have to experiment to see how much over exposure you need.
As sensors get better and better, the value of ETTR remains: smaller and denser sensors mean that ETTR can deliver near-DSLR quality in much smaller sensors, witness the Sony RX100, Sigma DP Merrill, Olympus OM-D E-M5 (and others).
Part of my reason for writing so heavily on this is for public “prior art” so no camera manufacture can screw the industry by patenting ETTR, claiming “new and non obvious”. I certainly think it is obvious, but patent examiners often grant ridiculous claims.
diglloyd blog - ETTR (Expose to the Right) — An Essential Skill to Master for Image Quality
ETTR is the single most important thing you can do for image quality with any digital camera besides Making a Sharp Image. Especially with small sensor cameras, ETTR makes images sharper by reducing noise that obscures fine image detail. diglloyd blog - ETTR (Expose to the Right) — An Essential Skill to Master for Image Quality