Topic : Photographic
Article 18 24 January 2007
16bit Files and Bird photography
Whenever you read about image editing or digital photography, the word ‘bits’ is mentioned, usually 8bits or 16bits, but also 24bits and 48bits. These ‘bits’ are units of colour information and give us a way to measure what we are working with. A pixel or picture element is the smallest visual element on a digital screen. A 1bit pixel has two values, on or off or in the binary system 1 or 0. When working with images with a 1bit pixel depth, the image would either be black or white.
This means that an 8bit pixel has greater pixel depth or bit depth and would display much more colour information. Remember a 1bit pixel has only two choices or levels of brightness either white or black so an 8bit pixel could have 2x2x2x2x2x2x2x2 or 256 levels of brightness, which means 2 to the power of 8 or 256 levels of brightness. An 8bit pixel uses one byte of memory. Therefore a 16bit pixel would use up twice the memory an 8bit pixel would use. The maths here would be the same for a 16bit image which means a 16bit pixel could have a possible 65,536 shades or levels of brightness (2 to the power of 16 = 65,536).
The greater the bit depth or pixel depth the image has the more colour information can be stored therefore the image should have a more accurate representation of the original.
Bits and Channels
An 8bit image would be an image with 256 shades, a greyscale image. As photographers we generally work with colour so our images are like a sandwich of 3 layers or channels, RED, GREEN and BLUE channels. Each of these RGB channels has 256 levels of brightness.
This means that an RGB image has 8bits per channel (3 x 8bits) and so is actually a 24bit image. And an RGB image with 16bits per channel (3 x 16bits) is a 48bit image.
So more bits per channel means more colour information and with bird photography that’s what we are after, realistic colour.
Why use 16Bits / Channel images?
Image editing is a destructive process. Just about any edit you do in Photoshop is going to cause some loss of image information to the image (levels of brightness), and this may happen when making any adjustments like curves or levels adjustments or even sharpening. Fortunately Photoshop CS2 allows just about all edits in 16bits/channel. So it would make sense to use as much information as possible to start with, then once all the edits have been done, convert the image to 8bits/channel if necessary, for example when it will be displayed on screen (email or web).
To try and illustrate this better I have made two copies of the same image, an 8bits/channel and a 16bits/channel image.
The same levels adjustment above was made to both images. The Black slider was moved to 20 and the White to 235 to try and increase the dynamic range of the image by spreading the remaining levels between 0 (black) and 256 (white). This made a significant difference to the image but at considerable cost.
By making this adjustment a total of 40 levels of colour information were lost from 256 levels, leaving only 216 in the 8bits/channel image. Just a few small edits to this image could cause noticeable banding and poor image quality.
In the 16bits/channel example, with the same levels adjustment, a total of 10240 levels are lost but this still leaves 55296 levels, far more to work with than 216 and still plenty for any other edits that might need to be done before any banding or real quality loss is noticed.
The information given in this article reflects my understanding on the topic. Birds are not easy subjects to photograph so use all the resources you can to get the best out of your picture. If you can get a high bit depth file out of your camera by shooting in RAW Mode, use it to make the most of your images.