Colour is one of the most confusing things in the digital world. Rule 1 is don't worry about it. Your camera is probably doing a pretty good job at capturing acurate colour and for most this will be sufficient (most cameras take pictures using 24 bit sRGB colour which is perfect for most applications). But if you want to learn about some of the various digital colour schemes, then read on:
If you go into some photo programs you'll see things like 8 or 16 bits/channel, RGB, CMYK, Paletted Color and other oddities. So what does it all mean.
24 bit RGB (aka 8 bits per channel RGB) - this is the most common colour model, three channels of colour (Red, Green and Blue) are used to produce all the colours the eye can perceive (yellow, brown, mauve, etc.). It's an additive model, full (255) values of Red, Green and Blue equals white. Each 8 bit channel can contain 256 colour graditions (28 = 256). Since there are 3 channels (Red, Green, Blue) this totals 24 bits.
To confuse the digital world, there are two main RGB types - one is known as sRGB and one is known as Adobe RGB. The sRGB standard is what most cameras, monitors and colour printers use. It was developed as a standard by Microsoft and others to most closely match the colour output of computer monitors, so that a photo displayed on one computer would look the same displayed on any other computer. The Adobe RGB system was developed with a wider colour range (gamut) to more closely match RGB colour space with CMYK (commercial printer) colour space.
For most people, including anyone working with 8 bits per channel, the sRGB system works best. The use of sRGB will most closely match what you will see in the camera to what you will see on the computer screen to what you will see on paper prints. Some pros working with 16 bit per channel RGB will opt for Adobe RGB and put up with the more complicated workflow for perhaps a slight colour advantage.
48 bit RGB (aka 16 bits per channel RGB) - this is a less common colour model in which each channel can contain 65,536 graditions of colour (216 = 65,536). Although this is much more than the human eye can perceive, pros like it since it provides more colour information in the photo, allowing for more accurate colour editing. Often images will be scanned and edited in 48 bit RGB and then the final result saved in 24 bit sRGB for computer display or printing.
96 bit RGB (aka 32 bits per channel RGB) - very rare and generally only supported by high end software this contains a huge amount of colour information (4,294,967,296 colour gradations per channel). This colour mode is supported by the TIF filetype, but not JPEG.
CMYK - Cyan-Magenta-Yellow-Key (Black) this is a colour model used for printing. Black (Key = Key Plate - a printing plate that was used to produce the darkest contrast colour, often black) is added as a separate colour rather than blending the three other colours to produce black. It is a subractive model (0 Cyan, Magenta and Yellow = White). It is usually best to allow a printing shop to do the RGB to CMYK conversion since there are several CMYK models, each one matching a particular printing process.
Paletted - this is colour restricted model, most commonly seen in specific file formats such as GIF, usually consisting of a colour palette of only 256 colours. Early web browsers could only display a particular paletted set of 256 colours and even today you may see this referred to in a photo program as "web safe" colours.
Greyscale - this is an 8 bit per channel system with only one colour, black. It provides 256 shades of grey.
Black & White - a 1 bit per channel system, either white or black. Good for text and line art, not much else.
To capture accurate colour, digital cameras use a function called "white balance." In a nutshell, it works to make sure that white is actually white in the photo. This is based on what is known as the "colour temperature" and the spectrum of the source light. For instance, normal sunlight has a colour temperature of 5,780K (degrees kelvin) and is full spectrum, a tungsten (incandescent) light bulb is about 3,200K with a restricted red end spectrum and a coolwhite fluorescent bulb is about 3,400K with a restricted green end spectrum. Your eye (or more accurately your brain) makes adjustments all the time for these differences. A digital camera tries to match this using its white balance. This works very well most of the time. It is a reason though why purists will shoot in RAW file format if the camera allows it - since you can adjust the white balance in a RAW photo, but you can't if the photo was taken in a "final" file format (i.e. JPEG, TIF). You can post-process a final format photo in photo software (using tools such as level adjustment and colour balance), but it's not as accurate as using a white balance adjustment.
No matter what colour model your photo is stored in, when viewed on a computer monitor, the luminosity (brightness) of the display is going to affect how you perceive those colours. For instance, the 24 bit RGB colour of red (255R, 0G, 0B) may appear to be a dark red on one screen and bright red on another, depending on the brightness of the viewing screen (and/or viewing angle when looking at an LCD screen).
There is a whole science to properly adjusting your computer screen to match actual RGB colour (see the brief mention of ICC profiles in the "Your Monitor" section on the Printing Page), but a quick and dirty method is to look at the filmstrip below:
Is this filmstrip black or can you see a message?
If you see a message, your monitor is too bright.
Turn down the brightness level until the filmstrip appears black.
For more information follow the copyright link:
Filmstrip image copyright: Timo Autiokari
Back to What Is A Digital Photo