Data Compression

​Data Compression

By Yefta Mandalika

Compression

A method that uses an algorithm to reduce the size of a file.

The benefits of having compression are:
1. Not as much storage space is needed to store the file.
2. It will take less time to transmit the file from one device to another.
3. It will be quicker to upload and download the file.
4. Not as much bandwidth is needed to transmit the file over the internet.

Types of Compression

There are two types of compression are
LOSSY and LOSSLESS Compression.

Lossy Compression

Lossy is a compression method that reduces the size of a file by permanently removing data.
This use compression algorithm that finds the unnecessary and redundant data in the file. This data is permanently removed from the file. This type of compression is mainly used on an image or sound file.

Lossy Compression

Unnecessary data in an image file that could be removed are colours that human eyes cannot distinguish . This means that an algorithm for an image file could reduce the size of the file by reducing the colour depth or the image resolution.

Unnecessary data in a sound file is also the one that human's ears cannot heard. Perceptual music shaping is a process that is used in lossy compression that removes sounds that are not normally heard by the human ear. The size of a sound file can also reduced are the bit rate and the bit depth,

Lossy Compression

Lossless Compression

Lossless is a compression method that reduces the size of a file by temporarily altering the data.

Lossless compression uses an algorithm to reduce the size of the file, without permanently removing any data. A common method that can be used to do this is by finding repeating data in the file and grouping this data together.

Lossless Compression

For example, consider the following message:

WHEN IT IS SNOWING HEAVILY LOOK OUTSIDE.
LOOK OUTSIDE IT IS SNOWING HEAVILY.

Excluding the spaces between the worlds and the full stops, the message has a total of 62 characters. 1 character requires 1 byte of storage, so 62 bytes of a storage would be needed to store this message.

These sentences consists of words that are mostly repeated. Instead of storing all 62 characters individually, the words and the positions at which they occur in the message can be stored in a lookup table. The

LOSSLESS Compression

WHEN IT IS SNOWING HEAVILY LOOK OUTSIDE.
LOOK OUTSIDE IT IS SNOWING HEAVILY.

​Word	​Position(s) in the message
​WHEN	​1
IT​	2, 10​
IS​	3, 11​
SNOWING​	4, 12​
HEAVILY​	5, 13​
LOOK​	6, 8​
OUTSIDE​	7, 9​

The lookup table needs 1 byte for each character in each word and 1 byte for each position the word occurs in the message.

Therefore, 33 bytes are needed to store the words and 13 bytes to store the positions, giving total of 46 bytes, this is much less than 62 bytes from original size. NO DATA HAS BEEN LOST!

To recreate the message, the computer simply retrieves the words and places them in the position allocated.

LOSSLESS Compression

A similar process can be applied to an image file with a lossless compression algorithm called run length encoding (RLE).
This algorithm will group together repeating pixels and store how many times they occur. For example, consider the simple image here:

If W is white, Y is yellow, R is red and G is green, RLE could compress the image into the following data:
12W, 3Y, 5W, 2Y, 1R, 2Y, 3R, 2Y, 3W, 2Y, 1R, 2Y, 5W, 3Y, 7W, 1G, 5W, 2G, 1W, 1G, 1W, 2G, 4W, 3G, 7W, 1G, 4W.

This means that the amount of data can be reduced as pixels are grouped together and the data for each pixel does not need to be stored. The resolution of the image is stored with the data, so that the computer knows the size to recreate it when the data is decompressed.