Code-excited linear prediction

About: Code-excited linear prediction is a research topic. Over the lifetime, 2025 publications have been published within this topic receiving 28633 citations. The topic is also known as: CELP.

An intrinsically reliable and fast algorithm to compute the line spectrum pairs (LSP) in low bit rate CELP coding

Abstract: The code excited linear prediction coder (CELP) makes it possible to synthesize good quality speech at low bit rates. In such a case, speech quality mainly depends on spectral envelope design accuracy. Different kinds of parameters belonging to the parametrical domain (linear prediction coefficients.

Coding distortion caused by a phase difference between the LP filter and its residual

Abstract: Several speech coding algorithms modify the time scale of the residual signal to facilitate efficient coding of pitch information. Time scaling, however, results in a phase difference between the coded residual signal and the time-variant linear prediction (LP) filter used for synthesis in the decoder. In this paper, we examine the coding distortion induced by this phase difference. Moreover, we show that it may cause audible artifacts to the synthesized speech even if lossless coding of all parameters is employed. These artifacts occur particularly at onsets when the frequency response of successive LP filters changes rapidly. A waveform interpolation coder is used to illustrate the effects of the phase mismatch.

A Multipulse-Based Forward Error Correction Technique for Robust CELP-Coded Speech Transmission Over Erasure Channels

Abstract: The widely used code-excited linear prediction (CELP) paradigm relies on a strong interframe dependency which renders CELP-based codecs vulnerable to packet loss. The use of long-term prediction (LTP) or adaptive codebooks (ACB) is the main source of interframe dependency in these codecs, since they employ the excitation from previous frames. After a frame erasure, previous excitation is unavailable and a desynchronization between the encoder and the decoder appears, causing an additional distortion which is propagated to the subsequent frames. In this paper, we propose a novel media-specific Forward Error Correction (FEC) technique which retrieves LTP-resynchronization with no additional delay at the cost of a very small bit of overhead. In particular, the proposed FEC code contains a multipulse signal which replaces the excitation of the previous frame (i.e., ACB memory) when this has been lost. This multipulse description of the previous excitation is optimized to minimize the perceptual error between the synthesized speech signal and the original one. To this end, we develop a multipulse formulation which includes the additional CELP processing and, in addition, can cope with the presence of advanced LTP filters and the usual subframe segmentation applied in modern codecs. Finally, a quantization scheme is proposed to encode pulse parameters. Objective and subjective quality tests applied to our proposal show that the propagation error due to LTP filter can practically be removed with a very little bandwidth increase.

Reducing search complexity for code-excited linear prediction (CELP) coding

Abstract: A code-excited linear prediction (CELP) coding method and code divide the residual signal into frequency bands. Codebooks provided for each band decrease in size with increasing band frequency. Reduction in codebook size with increasing frequency together with reduction in sampling rate with decreasing frequency provide reductions in codebook search complexity that allow real time implementation on digital signal processor chips.

Magnitude Spectrum Speech Hiding

Abstract: This paper presents a speech-in-speech hiding framework for the purpose of reducing the storage and transmission overhead in electronic voice mail applications, as well as for steganography applications of hiding secret speech messages for voice mail security. The technique used exploits the low-pass spectral properties of the Fourier magnitude of a host speech signal to embed another speech signal in the low-amplitude-high-frequency region of the host speech signal's spectral magnitude. Experimental evaluations on real male and female voice segments show that our technique is capable of hiding one speech message inside another host speech segment to produce a stego speech segment that is indistinguishable from the original host speech, while being able to extract the hidden speech message without any degradations in quality.