Bitmap colour depth

Because a bitmap image is really just a long string of 1s and 0s, it has no idea about the concept of colour.

We could store each colour as 24 bits (or 32 bits if we need transparency).
That would allow us to store pretty much any colour, but it does mean that images which only use a few colours would take up a lot more storage than they need to.

To fix this, we can store a bit depth, which will tell the program reading the image how many bits each pixel’s colour takes up.

Example: black and white images

If we store a black and white image, there are only two possible colours each pixel can have. We can represent a black-and-white image with just 1 bit per pixel (a 0 for black and a 1 for white, for example), and so our colour depth would be one bit.

Pre-set colour map

Some formats, such as Windows’ BMP bitmap format, have a hard-coded set of colours that all bitmaps will use.

For example, if we store an image in 4 bits, there are only 16 possible colours, and these colours are the same for all bitmap images of that format.

This isn’t very good if we want to store colours other than the pre-set map of the format’s colours.

Stored colour map

To fix this, we can use what I like to call a ‘stored colour map’. Instead of using the format’s default colour map, we can store a map (like a table) of the colour code, alongside the actual colour it makes up.

That way, if we have a specific colour used in our image, we can replicate exactly that colour, while also using fewer bits per pixel.

This map is stored in the metadata. It might look something like this if we visualise it:

Code	Colour
0	black
1	red
2	green
3	blue
4	yellow
…	…
15	white