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
Patent
07 Jul 2011
TL;DR: In this article, the authors propose a method to improve encoding quality when a frame to be encoded by a first encoding system does not continue when each frame is encoded by the CELP (Code-Excited Linear Prediction) encoding system or a second encoding system such as a frequency encoding system.
Abstract: PROBLEM TO BE SOLVED: To improve encoding quality when a frame to be encoded by a first encoding system does not continue when each frame is encoded by the first encoding system such as a CELP (Code-Excited Linear Prediction) encoding system or a second encoding system such as a frequency encoding system. SOLUTION: When a frame is encoded by the first encoding system or when a code encoded by the first encoding system is decoded, a decoded signal corresponding to a time-series signal at each point belonging to a past frame is linearly predicted and analyzed, and the obtained residual signal is used as a substitute for an excitation signal. COPYRIGHT: (C)2011,JPO&INPIT

2 citations

Proceedings ArticleDOI
21 Nov 2007
TL;DR: A bandwidth extension (BWE) algorithm for a low-bit-rate narrowband CELP coder is proposed using a spectral envelope sharing approach to develop a wideband speech coder, and it is demonstrated from a MUSHRA test with audio signals from four different music genres, that the BWE coder gives better quality than G.729EV layer 2.
Abstract: In this paper, we propose a bandwidth extension (BWE) algorithm for a low-bit-rate narrowband CELP coder using a spectral envelope sharing approach to develop a wideband speech coder. The developed wideband speech coder, referred to here as the BWE coder, is constructed using an embedded structure by adding an enhancement layer to the narrowband CELP coder. To minimize the bit-rate increase caused by the enhancement layer, the proposed BWE coder shares the spectral envelope and excitation parameters both with the narrowband CELP coder and the enhancement layer. In this paper, we choose G.729EV layer 2 as the baseline narrowband speech coder, and mel-frequency cepstral coefficients (MFCCs) are used to reconstruct the higher frequency components at the enhancement layer. By doing this, the bit-rate of the proposed BWE coder is found to be 12.7 kbit/s, just 0.7 kbit/s higher than that of G.729EV layer 2. It is also demonstrated from a MUSHRA test with audio signals from four different music genres, that the BWE coder gives better quality than G.729EV layer 2 and comparable quality to G.729EV layer 3, corresponding to an overall bit-rate reduction of 1.3 kbit/s.

2 citations

Proceedings ArticleDOI
23 Mar 1992
TL;DR: A real-time, TMS320C30 digital signal processor (DSP)-based implementation of a 4.3-kbit/s speech coder targeted for the E-TDMA half-rate digital cellular system is presented.
Abstract: A real-time, TMS320C30 digital signal processor (DSP)-based implementation of a 4.3-kbit/s speech coder targeted for the E-TDMA half-rate digital cellular system is presented. The speech coding algorithm is based on code excited linear prediction (CELP) technology. It features a 25-bit line spectral frequency split vector quantizer for encoding the short-term parameters, a 7-bit adaptive codebook, and a 9-bit fixed stochastic codebook. The short-term parameters, i.e., the line spectral frequency vector quantization indices, are updated every 20 ms and the excitation parameters i.e., the adaptive and fixed codebook indices and their gains, are updated every 8 ms. An overview is provided of the speech coding algorithm, the hardware platform architecture, the firmware development methodology, techniques used for optimizing the firmware of the real-time critical section of the algorithm, and speech quality monitoring methods employed during the firmware development. >

2 citations

Patent
16 May 1997
TL;DR: In this article, the authors proposed to simply switch the bit rate by providing multiple coding sections for the vector quantization of the time base wave-form, reference-inputting the quantization error of the (N-1)th stage for the coding of the Nth stage, and selecting the quantisation output of each stage to switch bit rate.
Abstract: PROBLEM TO BE SOLVED: To simply switch the bit rate by providing multiple coding sections for the vector quantization of the time base wave-form, reference-inputting the quantization error of the (N-1)th stage for the coding of the Nth stage, and selecting the quantization output of each stage to switch the bit rate SOLUTION: The second coding section 120 having a CELP coding structure is constituted of multiple vector quantization process sections, eg two coding sections 1201 , 1202 When the vector quantization by the closed loop search of the first stage is finished, the quantization error of the (N-1)th stage is used as the reference input for the quantization of the Nth stage (2<=N) The calculation quantity is reduced The case that both index outputs of two coding sections 1201 , 1202 are used and the case that only the output of the coding section 1201 of the first stage is used are switched, and the bit number can be simply switched

2 citations

Patent
22 Dec 1995
TL;DR: In this paper, a low-band evaluation voice is obtained by performing low frequency component extraction and down/up sampling of an input voice by a filter/down sampler 3, and the voice is evaluated with a synthesis voice generated using a linear predictive coefficient extracted by an extracter 4 and a down sampled code book 5 by an error evaluator 9.
Abstract: PURPOSE:To reduce an operation amount necessary for searching a code book in a voice encoding/decoding process using a CELP system and to maintain a certain extent of voice quality even in a less transmission information amount. CONSTITUTION:In a voice encoding device, a low band evaluation voice is obtained by performing low frequency component extraction and down/up sampling of an input voice by a filter/down sampler 3, and the voice is evaluated with a synthesis voice generated using a linear predictive coefficient extracted by a linear predictive coefficient extracter 4 and a down sampled code book 5 by an error evaluator 9. In a voice decoding device, a low band excitation signal is obtained by up sampling a decoded code vector by an interpolator, and a high band frequency component is generated from the signal by a high band adder by a PVE system to be added to a decoded excitation signal, and the voice is decoded by the signal and the linear predictive coefficient corresponding to the transmitted quantization data.

2 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