A Simulation-Based Comparison on Code Excited Linear Prediction (CELP) Coder at Different Bit Rates

High Quality Low Bitrate Voice Codec for Transmission Over Advanced Lte

“…1. In CELP, LP is applied to estimate the vocal tract parameters by predicting the current samples of speech frame as a weighted linear combination of previous samples [12] , [14] . Linear Prediction analysis (LP-analysis) is implemented on the framed voice signal s(n) to compute linear prediction coefficients â(i; m) of m order LP-analysis filter A(z) by applying autocorrelation estimate and Levinson-Durbin algorithm (L-D) [14] .…”

“…Linear Prediction analysis (LP-analysis) is implemented on the framed voice signal s(n) to compute linear prediction coefficients â(i; m) of m order LP-analysis filter A(z) by applying autocorrelation estimate and Levinson-Durbin algorithm (L-D) [14] . For encoding the parameters of the vocal tract with good quantization properties, the linear prediction coefficients â(I'm) are converted to the reflection coefficients k(i; m) , by applying inverse L-D recursion [12] . For achieving the analysis-by-synthesis concept, LP coefficients are used to form m th -order Linear Prediction synthesis (LP-synthesis) filter (z) with its transfer function is represented by [15] :…”

“…Another attractive approach in CELP is the use of a codebook that incorporates all the excitation sequences that are to be searched for the selection of the optimum sequence for encoding short voice segments in the encoder and only the selected sequence index is to be transmitted [9] . This method provides data bit compression, in addition to accomplishing highquality voice and bandwidth optimization [12] . During the paper's work time, many types of research in recent years are studied.…”

“…The study was showed that, at the values of perceptually weighted constant of 0.5, the reconstruction of the speech signal is better and audio quality and intelligibility are also maintained. Joshi et al, in 2018 [12] examined the performance analysis of CELP coding at 16 and 9.6 kbps bit rates with MSE and SNR performance metrics. The results showed that, at a lower value of c, the SNR be better and MSE is degraded.…”