Cours de 2 heures
| TP de 2 heures (ou 3)
|
The ``Coding, Transmission and Compression´´ is about the basics of information coding of simple data types: integers, floats, characters, strings, times and dates, error correction codes, redundancy and more complex data as the trees, the sounds and the picture.
After this, the quantity of information is defined and used to study the classics entropy encoding algorithms: Shanon-Fano, Huffman, Arithmetic. These algorithms are extended to a stream of dependent events and so the dynamic encoding is introduced.
Thereafter classic dictionnary compression methods are detailed: LZ77, LZ78, LZW, Block Sorting.
The bases of signal compression are introduced: noise, continuity, prediction, loss of quality, modeling. The full MPEG Layer 3 sound encoding is explained: DCT, psycho acoustic, RLE, encoding. The same thing is done with JPEG and MPEG.
Other pictures compression methods are presented: wavelets, Gaussian/Laplacian, JBIG. And compression and coding on other types: topology and geometry, TCPIP, HTML (WAP, JavaScript)
For all the methods it is explained if the method is adapted for real time and progressive transmission over a network.
The practical work is to create a compressor/decompressor of sounds and images without using any external libraries or tools.