Sensors linearity

Sensors are linear devices. If you double the amount of light, the sensor output will double, as long as the pixels are not full. If a pixel reaches full capacity, the output will be constant or clipped. Human vision is non-linear. If we double the light, the effect will be larger in low light conditions than in bright conditions, because our vision amplifies the shadows and compresses the highlight.

We can expose the sensor until the pixels are full and the brightest pixels output a value of 254, or we can halve the amount of light so that the brightest pixels output a value of 127. Either way, the result is that brightest stop uses up half of the 255 available tones and this is where human vision is least sensitive. The darkest stops, where human vision is more sensitive, can be described only by a few tones. This way, a very dark linear RAW image is created with a histogram skewed to the left.

For images to be more pleasing to the eyes, digital cameras apply a total curve to the linear row data. Applying a gamma correction of 1/2.2=0.45 will allocate more tones to the shadow area and fewer tones to the highlight areas in line with the characteristics of our vision. If we choose to work in a gamma 2.2 color space (RGB), our images  will appear perceptually uniform on a monitor. This way pasteurization can be avoided.

Cameras and row converters go beyond a gamma correction and apply more of a shaped curve to the data. This way a larger dynamic range is compressed so it can be represented on a monitor or print in a way that is pleasant to the human eye. Though, when using your camera some non linearity could appear in the darkest shadow and the brightest highlights. Also, some cameras preprocess the sensor data before the ADC.

Comments are closed.