I have in the past questioned whether shooting photos in RAW format adds anything, other than time spent hunched over the PC that is. I am though starting to be won over.
Here is a JPEG I took of my daughter Esther. She had just had an expensive hairdo and wanted a portrait. My ill-judged set-up had too much contrast and the sunlit garden out the window is well and truly overexposed. This is the JPEG as created in the camera:
Now this was the best I could get, starting with that JPEG and messing about with Photoshop. In practice all I could do was set the black point and emphasise the mid-tones a bit. There was no more detail to be had from the blown highlights:
Finally, the result of bringing the RAW file into Adobe Lightroom, reducing the exposure, using highlight recovery, adding fill light, adjusting the black point and applying some noise reduction:
The garden is still overexposed but a lot of detail has been rescued. Look also at the stripy curtains, the top of Esther’s arm, the details in the radiator just below the window.
Some of the benefit is lost because everything has been reduced to JPEG for web display, and fairly small JPEGs at that, but the RAW-based picture is far easier on the eye, without the same distraction and harshness caused by the large overexposed area.
In my earlier post I said I would not be convinced about RAW until I found an example where I could clearly obtain a better result with RAW than could ever have been achieved starting with the JPEG produced in the camera. Well this picture is certainly one such example.
I’m starting to see RAW as a “mini High Dynamic Range” image, ie RAW records more information than JPEG because light levels are recorded to 12 bits whereas JPEGs only hold 8 bits per channel for each pixel. A true HDR image made using say Photomatix combines several versions of the same picture, taken at varying exposures, to create a 32 bit composite. Anyone who has played with HDR images in say Photoshop CS3 will know how they look terrible because normal PC monitors lack the contrast range to display them properly, but as you adjust the luminance up and down different parts of the image take turns to display correctly. All the detail is in there, in the shadows and highlights alike, but you can’t display all that detail at once. Something similar is true of a RAW image, in a more modest way. It’s not quite the same; RAW is not explicitly” targeting” a wider dynamic range than JPEG but the differences in the way data is recorded do enable more data to be extracted particularly from the highlights.