第5章 无失真编码.pptVIP

Chapter 5  Lossless Source Coding 5.1 Lossless encoder General model of discrete, lossless (无失真）, non-memory source: How to code losslessly? Assuming that the statistical characteristics of the source can be ignored, it must satisfy the following condition. E.g. 5.1 Suppose we only have the first eight letters of English alphabet (A to H) in our vocabulary. The Fixed Length Code (FLC) for this set of letters would be A Variable Length Code (VLC) for the same set of letters can be Suppose we have to code the series of letters:” A BAD CAB”. The fixed length code and variable length code representations of the 7 letters are Instantaneous code (prefix code) (1) It is a kind of uniquely decodable code (2) In an unfixed length code, there is no one code being prefix to other codes. (3) When decoding, it does not need to refer to its following codes, and can make the judgment immediately, carrying out non-delay decoding. Let us have a look at Example 5.1 again. Now we consider another VLC for the first 8 letters of English alphabet: This second variable length code appears to be more efficient than the first one in terms of representation of the letters. Definition 5.1 A Prefix Code is one in which no codeword forms the prefix of any other codeword. Such codes are also called Uniquely Decodable or Instantaneous Codes. 5.2.1 Fixed length coding theorem If the equal length code is demanded to be uniquely decodable, then it must has: If N＝1，then： Conclusion：for unique decoding, each source symbol needs at least code symbols to be represented. When use dual-code symbol to code, r＝2，then： Conclusion：when use equal length dual-code, the limit code length of each source symbol is 5.2.2 Unfixed length (Variable length)  source coding Several concepts of code type(Eg.2.4.2) Non-singular code and singular code Uniquely decodable code Prefix code（instantaneous code, non-delay code） Kraft theorem