Sequentially Adaptive Prediction and Coding of Speech Signals

A new method of speech digitization called residual encoding is introduced, and its application to the speech digitization problem is studied. The residual encoding system is a form of differential pulse code modulation which utilizes both an adaptive quantizer and an adaptive predictor. The residual encoder differs from previous systems in two ways. First, a sequential estimation method is used to continuously update the predictor coefficients, and second, the predictor coefficients are not transmitted, but are extracted from the estimate of the speech signal at both the transmitter and receiver. No form of pitch extraction is employed. The residual encoding system with a Kalman filter or a stochastic approximation algorithm for identifying the predictor coefficients has produced good quality speech at a data rate of 16 kbit/s.     

Backward Adaptive Lattice and Transversal Predictors in ADPCM

Four different backward adaptive predictors and a fixed predictor are compared for use in an adaptive differential pulse code modulation (ADPCM) system for coding speech at 16 kilobits/second (kbits/s). For noise-free channels, the four adaptive predictors, a least squares lattice, a least mean square lattice, a Kalman transversal form, and a gradient transversal form, all exceed the fixed predictor performance as well as the performance of a continuously variable slope delta (CVSD) modulation system. For bit error rates (BER's) of 10^-3or greater, the transversal predictor performance falls below that of the fixed predictor and CVSD; however, the lattice structures maintain their performance advantage. The least squares lattice predictor has the best objective and subjective performance for both noiseless and noisy channels. All systems perform poorly for a BER of 10^-2. To extend the performance of ADPCM with a least squares lattice predictor down to a BER of 10^-2, the sampling rate is reduced and a selective coding scheme is devised. The resulting ADPCM system maintains excellent performance through a BER of 10^-2and outperforms CVSD for noise-free and noisy channels. The dynamic range, tandeming performance, and behavior for noisy inputs for the ADPCM system and CVSD are investigated.     

Sequential Adaptive Predictors for ADPCM Speech Encoders

The sequential gradient estimation predictor is compared in detail to the stochastic approximation predictor, and both are evaluated in an ADPCM codec. A switched predictor having two coefficients is then described for use in a DPCM-AQF codec. This predictor divides the range of the correlation coefficient of the speech signal into zones, and as the correlation coefficient changes zones, the predictor coefficients undergo a substantial modification. By this method the adaptation rate of the predictor is improved, particularly during transitions between unvoiced and voiced sounds.     

DPCM Coders with Adaptive Prediction for NTSC Composite TV Signals

New adaptive differential pulse code modulation (ADPCM) coders with adaptive prediction are proposed and compared with existing nonadaptive DPCM coders, for processing composite National Television System Commission (NTSC) television signals. Comparisons are based on quantative criteria as well as subjective evaluation of the processed still frames. The performance of the proposed predictors is shown to be independent of well-designed quantizers and better than existing predictors in such critical regions of the pictures as edges and contours.     

Sample Robbing in Predictive Speech Coders

In this paper, speech bit rate reduction by not transmitting a percentage of samples (i.e., robbing the coder of some samples) has been studied. The technique has been applied to predictive coders, namely differential PCM (DPCM) and adaptive DPCM (ADPCM) coders. A robbed sample is replaced by its estimate so that the prediction process in the feedback loop of the coders continues in a normal manner. After one period delay, when the next sample is decoded, the robbed sample is reestimated using delayed interpolation. Only periodic sample robbing has been considered, such as every fourth, every third, etc. The technique is particularly useful where graceful degradation is required under heavy loading conditions. The technique is found to be useful when the desired bit rate is 24 kbits/s or lower. The technique was evaluated by computer simulation using real-time speech inputs. Improvements of up to 3 dB in the case of a DPCM coder and of up to 1.5 dB in the case of an ADPCM coder have been achieved.     

Subjective Evaluation of Several Efficient Speech Coders

The voice quality of several 9.6 - 32 kbit/s coders is determined with an extensive set of subjective listening tests. Single encodings of μ255 PCM, adaptive differential PCM (ADPCM), subband coding (SBC), vocoder-driven adaptive transform coding (ATC), adaptive predictive coding (APC), and time domain harmonic scaling combined with SBC are compared in an idealized situation, that is, no added impairments. It is shown that single encodings of modest complexity 32 kbit/s coders such as ADPCM and SBC and more complex 24 kbit/s coders such as vocoder-driven ATC and APC offer quality nearly equivalent to 64 kbit/s μ255 PCM. However, these conclusions are drawn in the absence of a realistic telephone network where tandem encodings, delay limitations, and nonvoice signals exist. Tandem encodings of 64 kbit/s μ255 PCM, 32 kbit/s ADPCM, 16 kbit/s SBC, and 16 kbit/s APC are also evaluated. These 32 kbit/s and 16 kbit/s coders offer degraded tandem performance as compared to 64 kbit/s PCM, with the exception of synchronous tandeming of 32 kbit/s ADPCM with 64 kbit/s PCM where several encodings are subjectively equivalent to a single encoding of 32 kbit/s ADPCM.     

一种语音信号的线性局部自适应递推预测算法

语音产生过程包含非线性过程,传统的线性预测方法不能很好地解决这些非线性成份。局部线性预测是一种高精度的预测算法,但计算复杂度较大。为提高非线性预测的速度,提出了一种自适应递推局部线性预测算法．并设计算法的步骤,分析算法的复杂性。仿真结果表明,该算法比线性预测算法精度高,是一种有效的语音信号非线性预测方法。     

低比特率语声编码器的新发展

ITU-T建议G.729、G.729 AnnexA和G.723.1是国际电信联盟(ITU)最新颁布的3种适用于多媒体通信的低比特率线性预测语声编码器标准。文章介绍了语声编码器的比特率、复杂度、延迟和音质等性能指标的含义,并通过比较3种标准的新型声码器在算法和性能指标上的异同点,讨论了它们在多媒体通信中的不同应用。     

ITU G.729CS_ACELP语音编码系统

应用共轭结构代数码激励线性预测（CS－ACELP）算法的G．729协议是国际电信联盟（ITU）颁布的适用于多媒体通信的低比特率线性预测语音编码标准。介绍了G．729语音声码器编解码的基本结构，并对其性能指标：比特率、编解码的复杂度、时延和合成语音质量进行了分析。     

Detecting the Presence of Speech Using ADPCM Coding

When speech is coded using a differential pulse-code modulation system with an adaptive quantizer, the digital code words exhibit considerable variation among all quantization levels during both voiced and unvoiced speech intervals. However, because of limits on the range of step sizes, during silent intervals the code words vary only slightly among the smallest quantization steps. Based on this principle, a simple algorithm for locating the beginning and end of a speech utterance has been developed. This algorithm has been tested in computer simulations and has been constructed with standard integrated circuit technology.     

Parallelization of Context-Based Adaptive Binary Arithmetic Coders

This paper presents novel approaches to parallelize context-based adaptive binary arithmetic coders (CABACs). Two new parallelized CABACs (or PCABACs) are presented and the methods described. These coders are designed by modifying commonly used binary multiplication-free arithmetic coders. One utilizes linear approximation and simplifies the hardware by assuming that the less probable symbol probability is almost the same while performing the en/decoding (referred to as QL-coder). Another codec applies table lookup technique and achieves parallelism with a parallelized probability model (referred to as QT-coder). QL-coder is improved from the IBM Q-coder, and the QT-coder is improved from the CABAC used in H.264 video compression standard. Throughput, in both coders, is significantly increased after parallelization. A fast interval search method is also proposed.     

The Residual Encoder--An Improved ADPCM System for Speech Digitization

An improved system for speech digitization using adaptive differential pulse-code modulation (ADPCM) is described. The system uses an adaptive predictor, an adaptive quantizer, and a variable length source coding scheme to achieve a 4-5 dB increase in signal-to-noise ratio over previous ADPCM. The increase can be used to improve speech quality at moderate data rates on the order of 16 kbits/s or to retain the same quality and reduce the data rate to 9.6 kbits/s. The latter alternative permits the use of narrow-band channels. The implementation complexity is on the same order as other ADPCM systems.     

纯软件实时实现ADPCM语音压缩算法

本文提出了一个可在ＩＢＭ－ＰＣ／３８６或工作站上利用软件手段实时实现的语音压缩算法，压缩编码过程可在话筒和虚拟硬盘之间，反压缩译码过程可在硬盘和喇叭之间半双工实时进行，它可望在语音邮箱，多媒体语音系统中获得广泛的应用。     

基于改进RBF神经网络的ADPCM语音编码算法研究

以G．721ADPCM语音编码算法为研究对象,在语音编码的预测中引入神经网络模型来克服传统线性滤波方法中存在的不足,研究了基于RBF神经网络的ADPCM语音编码系统的结构。通过k均值聚类算法来确定RBF神经网络的中心和宽度,用最小二乘法确定RBF网络权值的方法改进了ADPCM语音编码算法。实验证明．其平均信噪比较原ADPCM编码算法有1-2dB的提高。     

基于分数阶滤波器的ADPCM预测误差信号处理

ADPCM音频压缩算法在G．721,G．723和G．726等相关数字语音通信协议中得到广泛应用。针对ADPCM中普遍存在的预测误差信号镜频干扰和高频噪声的问题,提出了基于分数阶积分器的误差信号频谱压缩滤波迭代算法和基于分数阶积分器的逆系统信号恢复迭代算法,为进一步提高音频压缩编码品质提供了新的方法,理论分析和仿真结果表明所提出方法的可行性。     

一种改进的语音信号非线性自适应预测编码方案

应用神经网络和Ｌｅｖｉｎｓｏｎ－Ｄｕｒｂｉｎ算法,本文提出一种改进的语音信号非线性自适应预测编码方案。用该方案实现了１６Ｋｂ／ｓ语音信号自适应预测编码器。实验结果表明,与原方案相比,本文提出的方案解码恢复后的语音质量有明显地改善。     

ITU-T Coders for Wideband, Superwideband, and Fullband Speech Communication

This section of the magazine presents recent algorithms developed by the ITU to provide high quality coding beyond traditional narrowband telephony. Speech coders can be characterized by their bit rate, quality, complexity, and delay. Typical applications fall into one of two categories, one-way and two-way. The first includes storage applications such as telephone answering systems, streaming, multimedia delivery, and push-to-talk calls. The second includes realtime communications such as two person phone calls and conference calls. In this latter category, if the delay is too large - exceeding 300 ms round-trip - humans have difficulty communicating, while for storage and playback operations delay is not a factor. The complexity of a speech coder is one of the main contributing factors to its cost and energy usage. Complexity is most often measured in terms of memory usage (both RAM and ROM) and the number of instructions executed per second. All applications are sensitive to cost, and many are sensitive to energy usage as well. The desired bit rate is determined by channel capacity or storage capacity, depending on the application.     

On the Performance of Speech Waveform Coders with Noise Spectral Shaping

The performance of speech waveform coders that incorporate a noise spectral shaping filter is discussed. The coders studied are adaptive pulse code modulation (APCM), adaptive differential PCM (ADPCM), and adaptive delta modulation (ADM). For APCM and ADPCM, the noise shaping filter is designed to minimize the C-message weighted quantization noise power. As for ADM, the noise shaping filter is used to move some portion of in-band noise outside the signal band. Simulation results show that the performance improvement of these waveform coders with spectral shaping is about 0.5-3 dB over the systems without noise shaping. Although this improvement is relatively small, the waveform coders with noise spectral shaping yield subjectively more pleasing and intelligible sound than those without it.     

Artificial Signals and Identification Methods to Evaluate the Quality of Speech Coders

Objective quality measures provide an economic and practical alternative to tedious and expensive subjective tests. Furthermore, some recently proposed measures predict quality values which correlate quite well with the subjective ones. The objective measures are particularly useful to test the influence of design parameters during the laboratory development of codec algorithms. In this paper the architecture of a measuring equipment to conduct objective quality measures on speech waveform coders is presented and discussed. Only the input and output signals from the coder are needed and an identification algorithm is used to separate the distortion effects introduced by the coder from other effects such as gains, time shifts, low-pass filtering, and so on. The choice of the most appropriate kind of signal for the identification and measurement steps is a problem also investigated. Particularly, results concerning mathematically defined speech-like signals are reported. The performance of the method is investigated under different conditions and the correlations with subjectively determined results are also reported for a number of ADPCM coders with widely varying parameters.