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 published on a yearly basis
Papers
More filters
24 Sep 2011
TL;DR: This paper presented a 11.2kbps speech coding algorithm, whose delay is 2.5ms and sampling rate is 8kHz, while a 20 dimensional CELP is used in this algorithm, while self-adaptive codebook search cascaded with algebra codebooks search are adopted in vector quantization.
Abstract: This paper presented a 11.2kbps speech coding algorithm, whose delay is 2.5ms and sampling rate is 8kHz. A 20 dimensional CELP is used in this algorithm, while self-adaptive codebook search cascaded with algebra codebook search are adopted in vector quantization. The PC simulation shows the synthetic speech quality after PESQ test is close to that of G.728 at 16kbps.
20 Sep 1995
TL;DR: In this article, a phase coding CELP (PC-CELP) was proposed to encode short-time phase characteristics of the LPC-excitation signal directly.
Abstract: In Code-Excited Linear Prediction (CELP) [l] coding at rates of 3kbps OT lower, efficient representation of the LPC-excitation signal is most important. However, at present, pitch periodicity and short-time phase characteristics of the LPC-excitation signal represented only by excitation Wdevectors are insuficient t o reproduce good quality speech. In this paper, we present a new CELP Coder, Phase Coding CELP (PC-CELP). In this coder, short-time phase characteristics of the LPCez&ation signal are extracted and encoded directly. Informal listening tests indicate that good quality synthetic speech is obtained at 2.5kbps.
TL;DR: Order-variable all-poles model according to instability of track complexity is provided and applies this model in Multi-pulses linear prediction speech coding to keep better speech quality on the base of decreasing coding rates.
Abstract: On the base of all-poles model, this paper provides order-variable all-poles model according to instability of track complexity and applies this model in Multi-pulses linear prediction speech coding. This method is simulated in Matlab and quality of synthesized speech is evaluated, order-variable model is founded to keep better speech quality on the base of decreasing coding rates.
22 May 1991
TL;DR: RELP schemes, which are based on a model of the vocal tract and the estimation of an optimum excitation function, are reviewed and refinements of this concept as vector sum excited linear prediction (VSELP) and post-filtering strategies lead to CELP codecs able to achieve a speech quality which is scored between fair and good in subjective MOS-terms.
Abstract: RELP schemes, which are based on a model of the vocal tract and the estimation of an optimum excitation function, are reviewed. In the RELP-S scheme the speech signal is interpreted as the result of filtering synthetic excitation functions by a filter which models the vocal tract. Two types of RELP-S schemes using different excitation generators are described. Refinements of this concept as vector sum excited linear prediction (VSELP) and post-filtering strategies lead to CELP codecs at rates between 8 kb/s and 4.8 kb/s able to achieve a speech quality which is scored between fair and good in subjective MOS-terms. >
01 Jan 2006
TL;DR: Objective and subjective measures show that the JPAS (Joint Position and Amplitude Search) multistage search technique proposed recently can yield to a very important reduction of coder computational load, without decreasing the perceived quality of wideband speech signals compared to other search techniques.
Abstract: In this paper, a low delay 16 kb/s wideband speech coder with a buffering delay of 1.25 ms is introduced. This coder is basically inspired from the G.728 LD-CELP standard for narrowband speech signals. Our main goal is to reduce the implementation complexity of our wideband G728 like coder, which is mainly due to the search of the optimal excitation (gain-shape) in codebook. For this reason an algebraic codebook is proposed and an exhaustive optimal search and suboptimal full position and Joint Position and Amplitude Search techniques are implemented. Objective and subjective measures on a large corpus of a testing speech database show that the JPAS (Joint Position and Amplitude Search) multistage search technique proposed recently can yield to a very important reduction of coder computational load, without decreasing the perceived quality of wideband speech signals compared to other search techniques.