The word "
Information" can be used to describe many things. Simplistically, what I meant is that if you have an image that contains Z amount of pixels and is an 8 bit image, then you'll have a certain amount of bits in the image. If you have a 16 bit image, the file will have twice that amount of bits.
The 8 bit image will have Z x 8 x 3, where Z is the number of pixels in the image, 8 is the number of bits per pixel and 3 is because there are 3 channels. Let's imagine there were 788,736 pixels (1024x768 image), there would be 18,874,368 bits in total needed to represent that image in 8 bit mode. This obviously doesn't account for bits needed for file headers/EXIF etc. It also doesn't take into account compression. If the image was 16 bit, then the image would be represented by 37,748,736 pixels - twice that of the 8 bit image.
You have to differentiate between the amount of information contained in bit stream, versus the range of colours/values that information can represent. A one bit data stream, has only one bit, but can represent 2 colours. Going to a 3 bit data stream, you now have three times the amount of information (i.e. 3 bits rather than 1) but you can represent a much larger range of colours - 8.
In information theory, information is quantified as the minimum average message length, in bits, that must be sent to communicate the true value of the signal to a recipient.
I hope that explains what I meant? P.S. I come from an electronics background (I have my own microelectronics design business:
IC Mask Design - IC Layout and Training Programs)
Quote:
Originally Posted by blumesan
Very good exposition; clear and to the point.
Don't know how you quantify information, so I wonder at the following statement:
I would have guessed that the figure would be much higher.
Cheers/Mike
|