Contents

1 History

• At the end of the 40’s, Claude E. Shannon (Bell Labs) and R.M. Fano (MIT) developed the Shannnon-Fano codec [1, 2].

2 Encoder

1. Sort the symbols using their probabilities.
2. Split the set of symbols into two subsets in a way in which the each subset have the same total probability.
3. Assign a different bit to each set.
4. Repeat the previous procedure to each subset until each subset has only one symbol.

2.1 Example

• Let’s use the following probabilistic model:

Using it, this is the Shannon-Fano coding:

2.2 Decoder

TO-DO.

References