SPEECH DATA COMPRESSION DEVICE AND METHOD THEREFOR, DECOMPRESSION DEVICE AND METHOD THEREFOR, TRANSMISSION SYSTEM FOR THE SPEECH DATA, AND PORTABLE TERMINAL DEVICE

摘要

PROBLEM TO BE SOLVED: To make natural sound highly compressible keeping high quality, and transmit it with good efficiency.;SOLUTION: A spectrum conversion means 2 samples PCM original sound of PCM sound source speech data at a prescribed rate, and divides them into the number of frames M in a prescribed sampling units, and then converts them into a frequency spectrum of prescribed frames. A spectrum extraction means 3 extracts N-pieces of spectra in order of prescribed weights to the frequency spectra of all the original sound. Each parameter such as a) large spectra, b) spectra of which the sum of the whole tone power is larger than that of the preceding frame at a certain ratio or more in frame units, c) a certain number of spectra in descending order in each frame, and d) an aural frequency characteristic of human ears, and an acoustic frequency characteristic of reproducing environment are set to an extraction parameter setting means 4. A bit stream preparation means 5 prepares a bit stream from a position and a size of the extracted spectrum, and transmits it as compressed data.;COPYRIGHT: (C)2002,JPO