Coding system for speech recognition - uses several successive ratios of extreme values for coding or correlation using resistance matrix

摘要

The method of extracting max and min. values of the waveform envelope of the speech signals includes establishing ratio values of adjacent extreme values varying against time and/or amplitude. From several successive ratio values, a coding is obtained. Groups of ratio values are obtained, by establishing the same repetitions of the values and/or the extremities of the envelope curves using the values, amplitudes and/or the time values. The same groups are counted or coded or combined with new groups, or only one of the similar groups is further processed. The groups are coded or correlated using a resistance matrix. Coding can result from the level of volume of words.