Topic
Enhanced Variable Rate Codec
About: Enhanced Variable Rate Codec is a research topic. Over the lifetime, 357 publications have been published within this topic receiving 4842 citations. The topic is also known as: EVRC.
Papers published on a yearly basis
Papers
More filters
31 Dec 2009
TL;DR: A candidate codec for ITU-T Embedded Variable Bit Rates (EV-VBR) standardization designed by Speech and Audio Signal Processing Laboratory (SASPL) of Beijing University of Technology is improved and the second sub-frame's spectral parameters are quantized additionally to improve the overall quantization precision.
Abstract: A candidate codec for ITU-T Embedded Variable Bit Rates (EV-VBR) standardization designed by Speech and Audio Signal Processing Laboratory (SASPL) of Beijing University of Technology is improved in this paper. In the improved codec, the second sub-frame's spectral parameters are quantized additionally to improve the overall quantization precision. Depth-first tree search algorithm replaces the full search and focus search in searching algebraic codebook. Transformation coding (TCX) process is redesigned and applied into the higher three layers which reduces the algorithm complexity dramatically. The objective mean opinion score (MOS) test results show that the improved codec achieves comparable speech quality as G.718 recommendation codec in most test items and has a lower algorithm delay.
1 citations
18 Dec 2006
TL;DR: The introduction of the proposed codec converts the variable-length symbols into fixed-length packets which can be hardwaredly and parallelly processing without waiting for any other information, which indicates the decoding process is context-free.
Abstract: A codec (encoder-decoder) design for interfacing variable-length and fixed-length data conversion is proposed in this paper. The poor memory efficiency of the variable-length compression approach can be avoided while its advantages can be preserved. The introduction of the proposed codec converts the variable-length symbols into fixed-length packets which can be hardwaredly and parallelly processing without waiting for any other information, which indicates the decoding process is context-free. The proposed codec encodes extra symbols in the redundant bits of the padding bits of the fixed-length packets. This novel encoding scheme relaxes the intrinsic poor bit rate of the traditional fixed-length data compression.
1 citations
Patent•
18 Dec 2004
TL;DR: In this paper, a method for sharing sound data of a terminal having a plurality of modems, and a method therefor are provided to process sound data to be shared by modem chips different in format, thereby efficiently using system resources.
Abstract: PURPOSE: A device for sharing sound data of a terminal having a plurality of modems, and a method therefor are provided to process sound data to be shared by modem chips different in format, thereby efficiently using system resources. CONSTITUTION: An AMR(Adaptive Multi-Rate) vocoder(220) receives an analog speaker signal output from a speaker output end of a synchronous EVRC(Enhanced Variable Rate Codec) vocoder(210) of a terminal through a microphone input end. The EVRC vocoder(210) receives an analog speaker signal output from a speaker output end of the AMR vocoder(220) through a microphone input end. An ADPCM(Adaptive Differential Pulse Code Modulation) vocoder(230) processes sound data of the AMR vocoder(220) and the EVRC vocoder(210).
1 citations
22 Jun 2006
TL;DR: This paper presents a reconfigurable architecture for the ACS block that enhances in area reduction of ACS block by almost 20% and gives significant reduction in time consumption.
Abstract: Today's modern mobile communications require minimum bandwidth, minimum loss, delay and better transmission of speech. This paved a way for the development of speech codec with low bit rates among which CELP codec inherit the merits of both waveform and source codec. The most time consuming and computational intensive process in modern speech codec is algebraic code book search (ACS). The 30% of process is done by this ACS in speech codec. This paper presents a reconfigurable architecture for the ACS block. This proposed architecture enhances in area reduction of ACS block by almost 20%. Further it gives significant reduction in time consumption
1 citations
13 May 2002
TL;DR: A new context-based error recovery technique for the CELP-based speech codec accomplishing recovery of erased frames, updating decoder state during erasure spells and reliable estimation of codec parameters in case of bit errors is proposed.
Abstract: GSM AMR speech codec being used for both Internet and mobile networks, robustness to both frame erasures and random bit errors assumes significance. This paper proposes a new context-based error recovery technique for the CELP-based speech codec accomplishing recovery of erased frames, updating decoder state during erasure spells and reliable estimation of codec parameters in case of bit errors. Previous error concealment techniques do not adequately make use of the context in which concealment is being done. The proposed error concealment technique is intended to retrieve and use contextual information for better performance. The method is solely receiver based, uses no look ahead, makes use of implicitly available codec parameters and buffers for parameter estimation and is hence computationally efficient. Segmental Itakura-Saito measure and MOS scores are used to compare the output speech quality of the proposed technique with those of the basic techniques as recommended by the standard.
1 citations