B1P2

Question 1

What is compression

Answer 1

reduce file sizes so they become more manageable to store, manipulate and transmit

Question 2

What is lossless compression

Answer 2

Also called noiseless or entropy coding
exact reconstruction of original source
Well known examples:
•	Lempel Ziv Welch coding
•	Huffman coding
Zip files
GIF, PNG and TIFF, final stage of JPEG

Question 3

What is lossy compression

Answer 3

Only approximate version of original source
Irreversible loss of information
MP3 and variants including AAC, HE-ACC, JPEG, JPEG2000, and MPEG coding family

Question 4

What is a fixed-length code

Answer 4

When a fixed number of symbols are encoded in to a fixed number of outputs

Source coding represents each message by a code word

Question 5

What is a code tree

Answer 5

A graphical representation of a procedure for the generation and decoding of instantaneous codes.

Question 6

What is meant by entropy

Answer 6

a measure of bits per symbol that can theoretically be achieved by source coding particular source

Question 7

How do you work out the efficiency

Answer 7

The efficiency E of a particular code can be quantifed by

E = H/l

Question 8

How is a Huffman code constructed

Answer 8

Symbols are ordered according to their probability with most probable symbol are the top and least at the bottom

Combine 2 lowest of each column

At each point where path splits, assign 0 to upper branch and 1 to lower

Question 9

How does a Huffman code compress files

Answer 9

New file will be in general be smaller than the original file, possibly a lot smaller, which compresses file
Sends a code table to decompress

Question 10

What is Run-length encoding (RLE)

Answer 10

A method of data compression based on sending numbers that indicate the lengths of runs of symbols of the same type.
It is not necessary to scan whole file before coding begins
Like Huffman coding, RLE has an overhead, but in the case of RLE the overhead is the single number associated with each run, whereas Huffman coding the overhead is the table for the whole file

Question 11

What is Lempel-Ziv-Welch coding

Answer 11

LZW coding works by detected patterns in data, assigning short codes to them, then using those codes when the same pattern appears again

LZW coding creates a coding table in which codewords represent sequences. The coding table has 4096 entries (212)

Question 12

What is meant by Pseudocode

Answer 12

– an informal description of an algorithm