Frequency-domain audio decoder system with entropy code mode switching

(57) Summary frequency domain audio coda chooses from different entropy coding modes according to the characteristic of an input stream. An input stream is divided into a frequency range according to a certain statistical standard drawn from the statistics analysis of the typical input or the actual input to code in detail. The entropy encoder optimized so that the data of the model of the range might be coded is assigned to each range. A mode selector applies the right entropy method to a different frequency range during coding and decryption. The boundary of division can be determined in advance and a decoder makes it possible to know implicitly which decryption method will be applied to coded data. Or it is possible by showing change of the coding mode to succession data to use the adaptation composition which attaches a flag to a boundary within an output stream. For example, the coda which created the dividing boundary which separates the frequency coefficient by which zero were mainly quantized from the coefficient by which non-zero were mainly quantized, next was optimized to each data is applicable. The whole more efficient processing is realized by associating the coding method according to the characteristic of input data. In fact, the number and frequency range of division change with data type which should be coded and decrypted.