Block 2 Part 2

Question 1

Compression

Answer 1

• applied to reduce file sizes so more manageable to store, manipulate and retransmit
• retains as much source content possible

Question 2

Different compression techniques

Answer 2

• AAC - advanced audio coding
• AMR - Adaptive Multi-Rate
• AVC - advanced video coding
• GIF - graphics interchange format
• HE-ACC - high-efficiency advanced audio coding
• JPEG - Joint Photographic Expert Group
• MP3 - MPEG audio layer 3
• MPEG - Moving Picture Experts Group
• PDF - portable document format
• PNG - portable network graphics
• TIFF - tagged image file format
• ZIP - named for compression algorithm

Question 3

Lossless compression

Answer 3

• also called noiseless or entropy coding
• enable an exact reconstruction of original source data from compressed representation
• well known examples are Lempel-Ziv-Welch (LZW) coding and Huffman coding
• ZIP, GIF, PNG, TIFF and final stage of JPEG use this

Question 4

Lossy compression

Answer 4

• irreversible loss of information
• only approx. version of original source can be reconstructed
• MP3, AAC, HE-ACC, JPEG and JPEG2000 for still picture coding and MPEG family use this

Question 5

Message

Answer 5

• consists of source symbol that is selected from predetermined set of possible symbols

Question 6

Source alphabet

Answer 6

• Complete set of possible source symbols

Question 7

Source coding

Answer 7

• represents each message by code word

- simplest source codes are fixed-length codes in which all code words contain same number of bits

Question 8

ASCII coding

Answer 8

• stands for American standard code for information interchange
• does not include accents or non-English characters
• UTF-8 (universal coded character set + transformation format - 8-bit) is another set used
• UTF-8 backward compatible with ASCII
• ASCII code word for each source symbol made up of its column number in binary followed by row number in binary

Question 9

Binary coded decimal

Answer 9

• each numerical value of decimal digit converted to binary

Question 10

Source extensions

Answer 10

• more digits used to represent more numbers or digits

- this lowers them number of bits per decimal needed

Question 11

Variable-length codes

Answer 11

• different numbers of bits used for different source symbols

Question 12

Uniquely decodable

Answer 12

• means that code can be decoded in only one way and can’t be mistaken for something else

Question 13

Instantaneously decodable

Answer 13

• as soon as any code word received, it can be decode immediately

Question 14

code trees

Answer 14

• instantaneous codes can be generated and decode using this

- rather like family tree

Question 15

Compressing files with Huffman code

Answer 15

• Huffman code used by finding probabilities of characters appearing in message
• In order to uncompress file Huffman code used needs to be sent with file
• adds more overhead

Question 16

Run length encoding (RLE)

Answer 16

• If symbols repeated you can add a number to represent how many symbols
• e.g. ccccc = c5

Question 17

Strength of RLE

Answer 17

• unlike Huffman coding, not necessary to scan entire file before coding
• overhead is only single number associated with each run unlike whole Huffman table
• Efficiency depends on amount of repetition in file

Question 18

Limitations of RLE

Answer 18

• requires lots of repetition in file to be useful