Folklore in source coding: information-spectrum approach

  • Authors:
  • Te Sun Han

  • Affiliations:
  • Graduate Sch. of Inf. Syst., Univ. of Electro-Commun., Tokyo

  • Venue:
  • IEEE Transactions on Information Theory
  • Year:
  • 2005

Quantified Score

Hi-index 754.84

Visualization

Abstract

Information theory has several traditional folklore problems about data compression or channel coding with reference to random number generation problems. Here, we focus on and reasonably formulate one of them from the viewpoint of information spectra. Specifically, we verify the validity of the folklore that the output from any source encoder working at the optimal coding rate with asymptotically vanishing probability of error looks like almost completely random