Showing papers on "Code-excited linear prediction published in 1981"
01 Apr 1981
TL;DR: This paper discusses a form of non-linear prediction, namely, the prediction of the phase of speech signals, based upon a new treatment of the classical speech production model within a short-time analysis/synthesis framework.
Abstract: Prediction plays a key role in many signal processing applications. Linear Prediction has, in particular, been extremely useful to the development of digital speech processing techniques and applications. There is however a growing need for improved forms of prediction. We discuss, in this paper, a form of non-linear prediction, namely, the prediction of the phase of speech signals. This study is conducted within a short-time analysis/synthesis framework and is based upon a new treatment of the classical speech production model. Experimental data are presented confirming the theoretical results. Finally the use of phase prediction to low-bit rate, high-quality coding applications is discussed.
7 citations
01 Apr 1981
TL;DR: An improved linear predictive algorithm that utilizes a synthesizer excitation function derived from the prediction residual that can be programmed to run in real-time on existing hardware is presented.
Abstract: We present an improved linear predictive algorithm that utilizes a synthesizer excitation function derived from the prediction residual The complexity of the algorithm is such that it can be programmed to run in real-time on existing hardware We describe in this paper the method we use to extract, evaluate, and quantize the excitation
3 citations
01 Apr 1981
TL;DR: Intelligibility test scores for a 1200 bit per second differential linear predictive coder demonstrate the potential for this technique even in channel error rates up to five percent.
Abstract: The concept of differential linear prediction is presented It is shown that for voiced speech, differential linear prediction results in a theoretical coding reduction of over twelve bits per frame compared to standard linear prediction Intelligibility test scores for a 1200 bit per second differential linear predictive coder demonstrate the potential for this technique even in channel error rates up to five percent
2 citations
01 Apr 1981
TL;DR: The use of two-dimensional predictive coding was shown to be superior to the earlier one-dimensional systems under the conditions tested and the superiority of the LAR domain for coding was verified.
Abstract: The purpose of this research was to investigate the use of two-dimensional predictive coding of area functions for data rate reduction in LPC-based speech coding systems. Predictive coding in the area function domain is a technique for further data rate reduction on the LPC parameters. Two objective speech distortion measures were used for the optimization of predictor parameters and for objectively assessing the speech quality. The superiority of the LAR domain for coding was verified. The use of two-dimensional predictive coding was shown to be superior to the earlier one-dimensional systems under the conditions tested. Two-dimensional interpolation using a minimum mean square error (MMSE) criterion for interpolation coefficients was shown to be less effective than standard linear interpolation under the conditions tested. However, a one-dimensional nonlinear interpolation using MMSE criterion coefficients was shown comparable to standard linear interpolation.