Lets take a photograph and scan it. We're going to add it to a document, proof it on an inkjet and send it for printing on a press. Wouldn't it be great if the colour could be relied upon at each stage, that you could look at your monitor and predict what the final product would look like. Well you can but only if you allow for the differences between the devices.
My eye can see billions of colours, a photograph is pretty good but will contain a good deal less, a scanner will reduce the colour gamut still further. A monitor is quite capable of producing a lot of colours, not as many as a photo but not bad, but inkjets and presses, they aren't great, their colour gamut is very limited compared to my eye.
Scanners and monitors are RGB devices, if you examine a monitor you will see red, green and blue pixels, their intensity controlled by the voltage placed across them. RGB devices start with no colour (dark screen) and have greatest signal in the white.
Inkjets and presses are CMYK devices, they start with a white background and add colour filters to reduce the light that reflects from the base. Their greatest signal is in the black. Just to be awkward my printers CMYK is different to your printers and they're both different to the press.
Imagine that there is a way of defining colour without reference to light or ink, a device-independent colour space. There is, it's called Lab. L is lightness, how white it is and the colours are arranged around a pair of axes labelled a & b. That gives us a standard to which we can convert RGB and CMYK.
A colour profile is just a database that holds a CMYK reference and it's Lab equivalent. Once you have the Lab reference you know what that colour is in RGB.
Then we can take a colour from the photograph and hold it throughout the workflow, but there are restrictions and compromises to be made because of the limited colour gamut on CMYK devices.
We will deal with that in a separate article.
