| 摘要 |
Method for delta-compacting language signals includes coding of samplings of language signal on basis of model of linear prediction, first one performs preliminary structuring of vectors in code book according to ratio of majoring of distances of those to given reference points, after that by means of analog-digital transformer one transforms language signal to set of digital samplings with coding frames of samplings obtained to vector of coefficients of linear prediction on basis of model of linear prediction, with transformation of coefficients of linear prediction to vector of linear spectral pairs, one performs buffering of group of frames, by Viterbi algorithms one performs re-calculation of function being evaluated for quantized values of frames, to communication channel for each frame one transfers difference between indices in code book from current to previous quantum value. |
