Sensor Linear Capture

The way that the camera sensor reacts to light falling onto it is very basic and known as ‘linear’. If the light that hits a sensor doubles in brightness then it will be captured as such. This is different to the way film and human eyes work. Eyes and film ’compress’ the light that they see so that double brightness doesn’t appear as great a difference as it really is. Cameras perform strong processing before we get to see the captured image and therefore we don’t notice that it is any different to the scene we’re actually shooting. If this processing wasn’t performed then the image would appear darker than it really is. To see the un-processed image you can shoot in RAW format and then process the image on a computer using a RAW converting software that allows linear processing. To process an image, the camera applies a ‘gamma correction curve’ which normalises the shot’s appearance.

I will attempt to simulate an unprocessed image by opening a JPEG file in Photoshop and applying a curve to darken the shot.

Original Image

I converted the image to 16 bits per channel because using the regular 8 bit mode could cause banding due to the strong adjustments I’m going to make.

Next I applied the curve below to my shot and the resulting image can be seen underneath.

This is what an un-processed image looks like before the camera applies a gamma correction curve.

Next I placed both images side-by-side and viewed their histograms.

It’s very obvious that the (simulated) un-processed image is a lot darker than the original jpeg. All of the tonal values are on the left side of the histogram showing that dark tones are very abundant. In the original (camera processed) shot, the histogram values are all spread along its length with a lot of values at the lighter end of the scale (right hand side).

I then produced a second curve in order to make the image appear back to normal. This is what the camera does when it processes a freshly taken shot. This is a Gamma Correction Curve.

The image below has my gamma correction curve applied to it. It is very similar in appearance to my original image (camera processed JPEG), although still slightly darker but that will be due to my curve not being quite right.

And finally I decided to compare the histograms from the original JPEG and my final gamma corrected image.

Original JPEG

Gamma Corrected with Curve

It is clearly visible that both histograms have a fair distribution of values along their entire scale. The image has a lot of dark tones and shadows thus the high amount of values on the left hand side. Overall the histograms are very similar however not identical as a camera is much more efficient at correctly altering an images gamma than I am. Fro this experiment I have learnt how my camera processes my shots to leave me with well exposed aesthetic images.