Abstract: Apparatus and method for reconstructing non-quantized adaptively transformed voice signals are shown to include noise shaping wherein the spectral envelope is scaled prior to generating bit allocation and energy substitution which is achieved after dequantization by generating the spectral envelope information for each block of transform coefficients based upon side information, generating transform coefficients which correspond to transform coefficients which were not de-quantized and for substituting the generated transform coefficients into said blocks; and transforming said blocks of de-quantized transform coefficients and generated transform coefficients from said transform domain into said time domain.

Abstract: A method and apparatus for removing the periodicity from a speech signal in a transform coder prior to the quantization of the speech signal, which speech signal is a sampled time domain speech signal composed of information samples, the transform coder sequenctially segregating the speech signal into blocks of information samples, is shown to include apparatus and method for determining the pitch in each of the sample blocks, determininig a long term predetermined parameter (LTP) for each of the blocks based on the pitch determined for each block, calculating a periodicity value for each sample in the block wherein the calculation of the periodicity value is based upon the pitch and the long term predictor parameter, generating a revised block of difference samples by subtracting the periodically value from the corresponding sample, and performing adaptive transform coding on each of the difference blocks.

Abstract: A transform coder operates on a sampled speech signal transformed from the time domain to a frequency domain to develop pitch information in relation to a given speech signal. The coder segregates groups of information samples into blocks, transforms each block of samples, and generates an auto-correlation function of the transformed signal for each block. Next, the coder determines the pitch period and pitch gain from the auto-correlation function, and determines the striation magnitude and energy from the pitch period and pitch gain. Then a reference pitch model including a number of data points is retrieved from data memory. A striation scaling factor is generated in response to the striation magnitude and energy, and is multiplied by each of the retrieved data points to adaptively generate a pitch model. Finally, the adaptively determined model is sampled to establish the pitch information.

Abstract: Adaptive transform coding of a speech signal by a single digital signal processing chip is performed at low bit rates with reduced quantization noise and distortion. A windowed speech signal is transformed and quantized. New processes are shown for generating envelope information as well as bit allocation which control quantization. The quantized signal and necessary side information are formatted for transmission and subsequent decoding.

Abstract: A linear predictive echo canceller is integrated with a RELP vocoder in a two-way communications network, in which a received digitized speech signal synthesized by the RELP vocoder from a residual signal received with linear prediction coefficients is echoed onto a transmit channel and combined with a transmitted digitized speech input signal that is to be analyzed by a RELP vocoder for transmission. The echo canceller cancels the echoed signal from the transmit channel by generating a foreground transmit signal by subtracting a foreground estimate of the echoed signal from the combined signal; and providing the foreground transmit signal on the transmit channel for ananlysis by the RELP vocoder.