| 摘要 |
PURPOSE:To reduce the storing capacity of a pitch predictor and, at the same time, to produce reproduced voice signals irrespective of the connecting order of the pitch predictor and a liner predictor on an encoding side by contriving a constitution which uses the reproduced voice signals only when the reproduced signals are produced. CONSTITUTION:When the outputting code sequence of a voice encoding system which encodes voice signals by means of a pitch predictor 3, linear predictor 2, and voice source parameters corresponding to a prediction residue is decoded, the predictors 3 and 2 are cascade-connected. In addition, only reproduced voice signals are used when calculating the reproduced voice signals. Therefore, the reproduced voice signals can be produced irrespective of the connecting order of the pitch and linear predictors on the encoding side and, at the same time, the storing capacity can be reduced. |
