Speech Signal Coding Using Forward Adaptive Quantization and Simple Transform Coding

Abstract

The paper proposes a novel speech signal coding scheme that implements a simple transform coding and forward adaptive quantization. The proposed scheme is adapted to the input signal variance, providing highly efficient bandwidth usage, whereas implemented transform coding provides sub-sequences with more predictable signal characteristics, so that more suitable signal processing can be performed. The aforementioned transform coding precedes adaptive quantization, providing additional compression. The objective quality measure used for system performance estimation is SQNR (signal-to-quantization-noise ratio), which represents a standard measure for lossy coding types. The influence of transform coding is discussed by comparing the obtained results with the corresponding one achieved by applying only the same adaptive quantization. Furthermore, the comparison with system performance of PCM (pulse-code modulation) coding system confirms that the proposed coding scheme has a lot of potential for further implementation, since that the proposed system ensures SQNR gain up to 4.0983 [dB] for various values of system parameters.

Simple Speech Transform Coding Scheme using Forward Adaptive Quantization for Discrete Input Signal

“…Next, simple transforms t 1 and t 2 are applied to form two sub-sequences y 1 and y 2 , creating two independent signals with more predictable characteristics. These trans-forms are defined similarly as Hadamard transform for a group of two samples and their form is [32], [30]: based coding step is to provide additional compression before adaptive quantization [32], [30]. After that, speech signal is divided into frames whose size is 240 samples, which are further adapted to the variance framewise [5], [27], [21].…”

“…Next, simple transforms t1 and t2 are applied to form two subsequences y1 and y2, creating two independent signals with more predictable characteristics. These transforms are defined similarly as Hadamard transform for a group of two samples and their form is [32], [30]: ( ) 2 :…”

“…where xn and xn+1 are neighboring samples of the input signal, while y1 and y2 represent samples of transformed signal. These transformed signals have variances 12 σ and 22 σ , respectively, whose values depend on the discrete input signal variance σx 2 and correlation coefficient ρ [30]:…”

“…(2) where x n and x n+1 are neighboring samples of the input signal, while y 1 and y 2 represent samples of transformed signal. These transformed signals have variances σ 12 and σ 22 , respectively, whose values depend on the discrete input signal variance σ x 2 and correlation coefficient ρ [30]: adaptation technique. The aim of such transformbased coding step is to provide additional compression before adaptive quantization [32], [30].…”

“…These transformed signals have variances σ 12 and σ 22 , respectively, whose values depend on the discrete input signal variance σ x 2 and correlation coefficient ρ [30]: adaptation technique. The aim of such transformbased coding step is to provide additional compression before adaptive quantization [32], [30]. After that, speech signal is divided into frames whose size is 240 samples, which are further adapted to the variance framewise [5], [27], [21].…”