Page 57 - DCAP303_MULTIMEDIA_SYSTEMS

Unit 4: Image

As it can be seen from this table, large 24 bit images will result in very large files. This is why notes
compression becomes important. There are a large number of file formats used for storing
compressed bitmaps from the trival to the very complicated. The complicated formats exist because
of the very large bitmap files that would exist, if compression was not used. There are two broad
categories of compressed file format, those which are lossless (retain the bitmaps perfectly) and
those which are lossy. The following shows the main hierarchy of compression techniques (See
Figure 4.15).

figure 4.15: Hierarchy of Compression techniques

The crudest way of reducing the size of bitmap files is to reduce the colour information; this is
called bit reduction or quantization. For example, one could convert 24 bit bitmaps to 8 bit indexed
bitmaps using dithering to simulate the lost colours. The most common lossy format by far is JPEG,
a description of how it works is well outside the scope of this discussion. Its main advantage is
that it can offer vastly better compression ratios than the lossless formats. For example, consider
the following bitmap the original of which is 500 × 350 pixels at 24 bit colour. Using the formula
given earlier the uncompressed file size is 500 × 350 × 24 / 8 / 1024 = 513 K (See Figure 4.16)

figure 4.16: Uncompressed file

Saved in greyscale (bit depth reduction) the file is 171 K (3 times smaller), saved and compressed
using RLE, it is 388 K (75% of the original), saved using LZW compression, it is 188K (36% of the
original), saved as JPEG, it is 30 K (a compression ratio of 17 : 1). The following is a description of
the simplest lossless compression technique called run length encoding (RLE) that is used with
good effect for bitmaps with only a few colours. Consider the following small, 17 × 10 pixel, and
8 bit image (See Figure 4.17).

figure 4.17: 8 Bit image

LoveLy professionaL University 51