Digital Light & Color

I almost understad why it may work, but implementation sounds cumbersome. Did anyone used this technique in practice?

http://www.malch.com/nikon/UniWB.html

Maciej, Iliah Borg has written numerous forum responses concerning UniWB and he has developed curves for several Nikon models. You might try a search based on his name plus UniWB for user experience.

No experience with UniWB. Perhaps almost as good and easier: buy an exposure meter. Experiment to find a formula, easy to calculate in the field, that relates the meter reading to a roughly optimal exposure by the camera.

Here is a more readable article on the topic:

http://www.guillermoluijk.com/tutorial/ ... dex_en.htm

A side question: how do Maker's Notes EXIF RGGB levels translate to the channel 'multipliers'?

Maker notes are proprietary and their formats are generally not published, so there is no uniform answer to your question or even the assurance that these mutipliers are necessarily part of all makers notes. However you can find results of reverse engineering of various camera maker notes on the web or built into some third party exif programs.

Kiril

Here is a discussion on the EXIF WB multipliers (with the contributions from Guillermo and I think our very own Elie...):

http://photography-on-the.net/forum/sho ... p?t=854547

Forgot to ask (and I don't seem to be able to find appropriate reference): when adjusting WB in RAW processors, say in 'As Shot' option, the multipliers reported in EXIF must be converted to Colour Temperature and Tint couple somehow (and vice versa: the Colour Temperature and Tint set by the user have to be converted to the RGGB scaling factors?). What's the conversion equations/algorithm?

Just one more article on the UniWB topic. It's all theoretically interesting, but is it of practical value: is equating RAW channel multipliers going to be such a panacea for optimizing the exposure when shooting RAW? Do people have any experience with tweaking RAW channel multipliers?

Cheers!

http://www.libraw.org/articles/white-ba ... meras.html

Does the camera's metering mode [multi-zone, average, spot] have an impact on how far one can "expose to the right" the RAW histogram without channel clipping?

I'm still learning, but I think this is all about interpreting the histograms. There are at least two issues:

1) the in-camera life histogram likely shows the luminosity (mine does) - a weighted average of gamma-transformed RGB values. It shows clipping when all 3 RGB channels are saturated (pure white), but not when only one or two are. The UniWB can't help here, I think. The individual R+G+B histograms in review would show clipping of individual channels.

2) the in-camera histograms (regardless whether it's luminosity or the individual R+G+B channels) are based on in-camera processed, WB-corrected JPGs. They show clipping of the JPG values, not necessarily of the RAW values. Because the WB multiplies the original R and B RAW values depending on the WB chosen to produce a good-looking JPEG, those two channels may be shown pessimistically clipped when looking at the in-camera R+G+B histogram, even though the RAW R and B may not be saturated and can be handled properly in RAW post-processing. UniWB is a trick to set all the multipliers to 1, which produces an ugly-looking JPG but does not affect the RAW at all, and thereby make the in-camera R+G+B histograms reflect the actual RAW values distribution (allowing to judge the actual channel clipping as defined by maxed-out RAW values while adjusting the exposure).

Does it make sense?