Real-time expert system for compressing/decompressing digitized analog signals without distortion.

摘要

Real-time expert system for compressing/decompressing digitized analog signals without distortion. The expert system is based on the reordering of the information available after signal digitization. For this, the set of digital codes originating from discretization is converted into another set of whole numbers each obtained as the difference of one digital code from the previous one, assigning the value 0 to the first sample. (The recovery of original data is ensured by saving the digital code of the initial sample). The compression itself is carried out on the integers resulting from the conversion. They are coded according to a delayed method that takes advantage of the fact that 98% of the resulting distribution is usually concentrated around 1% of the possible values, which always correspond to numbers around 0. (The figure shows this fact for the conversion resulting from a 12-bit digitized signal). The bit streams thus obtained can be processed (storage, transmission, etc.) with the recovery of the initial data being assured. This expert system is of application in all those areas that demand data compression. Rates of over 80% have been achieved. IMAGE