Very true - as an evaluation of the entire scene, histograms don't lie .... but they can certainly mislead. The histogram is always a valuable guideline and good starting point for determining what you want to do with an image. But if one blindly trusts it, that can lead to some difficulties.
My main work is weddings and social events. Many of my images involve a white bridal gown against a dark larger background, or dark suits/tuxedos against dim backgrounds with light colored faces. A histogram will "judge" the general value of these scenes. But in everyone of these I will need to go in and either dodge and burn, or create exposure adjustment layers to achieve the desired pleasing results.
Canon ETTL with and without flash does a fairly decent job once you understand it's foibles and peculiarities, and it does generate some decent histogram data. But that data is not the end all/be all of every image. The histogram never lies, but it can consider factors that are not your main image goal
.
Btw - nice B/w interpretation. I prefer the second one better. IMHO it has a very nice tonal rendering and contrast range, which suits the subject better than a sepia representation. But that's just my opinion
- Stu
Irreverent wrote in post #5863103
Um, no they wouldn't. Histograms don't lie. If you look at the histograms for the images, they are pretty heavily under-exposed. Here's how far I managed to push them without clipping any highlights.