HDR / Discussions

I think not all our Ipernity friends know exactly what's HDR.
HDR stands for High Dynamic Range.
The problems: film and sensors haven't the same ability of the human eye to record high luminosity differencies. A typical JPG image has a range of 256 different tones for each of the three R(ed)G(reen)B(lue) colour components. Digital sensors can have 16 to 256 times this range. Human eye has greater range. So, in a typical scene with a dark subject against a light background, in a traditional LDR shot we have to choose if we want details in the highlights zone or in the shadows zone.
The solutions: film and modern sensors capture a much higher range, but this only partially solves our dilemma --> the standard image formats are 8-bits ones (256 tones x component).

First step: capture a wider range --> the RAW formats have 12 to 16 bits per component;

second step (optional): capture a more wider range --> use the bracketing: mount the camera on a tripod and capture the scene on different shots, changing the exposure value (3 to 15 different exposures are usually employed)

third step: combine the different exposures with specialised softwares (Photomatix and QtPfsGUI are the more common) with a technique called tone-mapping: use details from the lighter exposures for the shadows zone and from the darker exposures for the highlights zone (if we don't make the second step, we have a pseudo-HDR and data from teh same exposure are combined to obtain the pseudoHDR shot).

A visual example:

- original shot (directly converted to JPG):
"Roar" by Roberto Ballerini [?]



- (pseudo)HDR shot (tone-mapped and converted to JPG):
"Roar (HDR)" by Roberto Ballerini [?]



As you can see, there is a lot more detail in the lion's face, particularly in the ears, but the sky looks too saturated and a little bit unnatural.