Jan 31
1. Go over assigned exercises in the text, including the one using Shannon-Fanon coding. (Look that up in wikipedia if no one is ready to describe it.)
2. Finish discussion of information entropy (and markov-1, markov-2, etc approximations), including how (explicitly) to calculate it, using someone's code, either one of yours or mine that I linked to last week.
Move on to chapter 4 material in "Codes" text : compression.
Two overall compression types: lossless, lossy. We'll start with the first.
Discuss 3 variable length output compression codes:
- Shannon-Fano
- Huffman
- Arithmetic
Mention
Dr Ross Compression Crypt : copyright and patents loom large in the story of which of these algorithms are typically used.
Describe LZW (someone who's done it before, or refer to wikipedia).
Last of lossless : Burrows-Wheeler. Assign some reading for next time.
Question: which of the above techniques are used in the common archive formats (.zip, .gz, .bz2, ...) ?
Coming: lossy formats :
- images (jpeg ...). Quick quiz: is .png lossy or lossless?
- sound/video (mp3 ogg ...). Quick quiz: name a lossless audio encoding.
Start that discussion if there's time.
