The Myth of the Unmanipulated Image

[chakalakasp]

A rather quick and somewhat provocative article -- and one that I fully agree with. While one can certainly defensively hold a position that there is a line to be drawn where manipulation takes a photograph beyond truth, it's very difficult to maintain a position in which only images farted directly out of a camera qualify as "real" photographs.

The Myth of the Unmanipulated Image | BH Insights

I recently went to a talk put on by a Natty Geo photographer who went to great lengths to describe how he never digitally corrected any of his images (or cropped them), and how everything we saw on the screen was right out of the camera. His audience of mostly college students ate this crap up. As someone who used to work pre-press, it astounds me how uninformed some top-notch first-in-their-field photographers can be of how an image goes from their developed slide to a magazine spread or a digital projector screen or a digital image on their laptop screen. They get how a camera uses light to create an image on film or a sensor, but after that, they close their eyes and everything is Magic to them That Must Not Be Spoken About. Unless you are holding the undeveloped slide in your hand, the image has been processed. Scanning a slide or a neg requires further processing it by definition; how it is processed is up to the scanning technician and the scanning hardware/software. Getting images to look good even on good a Flexo press printing on heavy slick paper requires quite a bit of post processing -- it's just done by prepress folk, not photographers (who merrily go on their way thinking that their 'virgin' image remains unsullied).

 

[KmH]

Yes, there is all of that and more.

I agree with the B&H article but the writer didn't quite get it right with his 101100100111010111101000011101010 analogy.

The image starts as an analog voltage stored by each pixel. The amplitude of that voltage is determined by the amount of light that strikes each individual pixel. The apmplitude of the voltage is then amplified by an amount determined by the camera's ISO setting.

That amplified voltage then gets converted to a digital number so it's more like 101011001001, 110010100001, 001010100011, for a 12-bit Raw data file, each 12 bit number representing the now amplified voltage recorded at a single pixel.

A Raw data file is actually a grayscale image.The image sensor does not record the colors of a scene. The colors are inferred from a known color array filter that is in front of the image sensor. Most digital cameras use a Bayer filter: Bayer filter - Wikipedia, the free encyclopedia.

There is an algorithm (de-mosaicing) that the pixel digital data is run through that infers from the luminosity each of the 4 Bayer array pixels 'saw'.

 

[spacefuzz]

Cool article, I agree.