scispace - formally typeset
Search or ask a question
Topic

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.


Papers
More filters
Proceedings ArticleDOI
Di Gao1, Xiaoqun Zhao1
21 Jun 2013
TL;DR: A new 600 bps quantization method based on the MELP parameters is proposed, which contains methods like inter-frame prediction, linear interpolation and so on and though the bit rate is only a quarter of the original M ELP algorithm, the scheme could also synthesize a reconstruction speech which is understandable.
Abstract: The MELP is the U.S. Federal Standard operated at 2.4kbps, it could achieve equal to or better than the performance of the 4.8kbps CELP vocoder. However, for channels like underwater acoustic channel, the condition is very bad that even the 2.4kbps MELP data could not be transferred. A new 600 bps quantization method based on the MELP parameters is proposed. The new scheme contains methods like inter-frame prediction, linear interpolation and so on. Though the bit rate is only a quarter of the original MELP algorithm, the scheme could also synthesize a reconstruction speech which is understandable.

3 citations

Patent
01 Mar 2010
TL;DR: In this paper, a speech signal is analyzed by speech sample, and parameters of CELP (code-excited linear prediction) model, that is an LP filter coefficient, and adaptive and fixed code book indexes and a gain are extracted.
Abstract: PROBLEM TO BE SOLVED: To provide an optimal speech codec for solving the problem that, it is often difficult to suitably make a model of noise in a conventional method for encoding noise, resulting in undesirable interrupt, discontinuity, and during conversation, and that background noise is not suitably encoded in a speech encoder of analysis by synthesis, like a conventional code-excited linear predictive encoder particularly in a reduced bit rate. SOLUTION: A speech signal is analyzed by speech sample, and parameters of CELP (code-excited linear prediction) model, that is an LP filter coefficient, and adaptive and fixed code book indexes and a gain are extracted. These parameters are encoded and transmitted. LP analysis are performed twice by frame in Block 239, and only a single LP parameter set is converted to a quantized vector by using line spectrum frequency (LSF) and predictive multi-stage quantization (PMVQ). COPYRIGHT: (C)2010,JPO&INPIT

3 citations

Proceedings ArticleDOI
03 Nov 1996
TL;DR: To improve speech intelligibility and quality for coding, an array preprocessor is proposed to enhance noisy speech and preliminary results show that the array gain is transparent to CELP but not necessarily to LPC-10.
Abstract: To improve speech intelligibility and quality for coding, an array preprocessor is proposed to enhance noisy speech. The microphone array data is processed with a robust quadratically and linearly constrained minimum variance algorithm. The processor nulls noise and/or interference power from arbitrary spatial and frequency locations and maintains speech from desired locations, thus increasing the SNR for subsequent coding. The array and coder (LPC-10, CELP) structure is studied in terms of array processing gain and overall spectral distortion measurement. Preliminary results show that the array gain is transparent to CELP but not necessarily to LPC-10.

3 citations

Proceedings ArticleDOI
01 Apr 1985
TL;DR: The strategy proposed here selects the all-pole parameters to concentrate the model excitation in a finite number of locations to produce a maximally pulse-like residual as a result of theall-pole parameter estimation.
Abstract: Multiple pulse excited linear predictive coding (MPLPC) has recently received a great deal of attention in the literature as an attractive means of speech coding at data rates below 10 Kbits/second. The existing approaches to MPLPC analysis arrive at the parameters for an all-pole model by minimizing the mean squared modeling error before attempting to find a set of pulses to excite the model. The strategy proposed here selects the all-pole parameters to concentrate the model excitation in a finite number of locations. The goal is then to produce a maximally pulse-like residual as a result of the all-pole parameter estimation.

3 citations


Network Information
Related Topics (5)
Decoding methods
65.7K papers, 900K citations
83% related
Data compression
43.6K papers, 756.5K citations
83% related
Signal processing
73.4K papers, 983.5K citations
83% related
Feature vector
48.8K papers, 954.4K citations
80% related
Feature extraction
111.8K papers, 2.1M citations
79% related
Performance
Metrics
No. of papers in the topic in previous years
YearPapers
20226
20213
20207
201915
201810
201713