This paper describes modeling and simulation of candidate AM subcarrier systems to determine their relative performance. Based on the simulation results, a prototype system is designed that was successfully tested on the air. A data rate of 200 bits per second was achieved without introducing main channel audio distortion; the digital data was recovered in an automobile over 200 miles from the transmitter site well beyond the Continental Divide through the Colorado Rocky Mountains. For the simulation study, the primary goals were to determine the level of interference between the digital baseband spectrum and the main channel audio spectrum, and to identify effective subcarrier modulation schemes. In early systems, the phase shifts modulating the main carrier were limited to +/-15 degrees to avoid significant audio program distortion during AM detection. This paper explores the possibility of applying other digital modulation techniques including BPSK, QPSK, GMSK and QAM. Sample speech segments are used to create the AM portion of the composite signal. Effects of the secondary RBDS signal on the voice program output signal-to-noise ratio are assessed. Conversely, the effect of the AM sidebands on the RBDS bit-error rate is measured. Finally, a prototype system is described using off the shelf components to demonstrate this technology on the air. Low rate digital data is digitally modulated onto the main carrier of a commercial AM radio station. A conventional car radio is modified to extract the soft bit stream, which is subsequently processed in real time within a laptop computer hosting a DSP board. The performance of this system during field trials, transmitting from commercial AM station KTLK in Denver, Colorado, is summarised

Output-based speech quality (OBQ) refers to an objective speech quality measure that uses only received speech without access to the input speech record. This paper proposes two new OBQ measures and evaluates their performance. Perceptual linear prediction (PLP) coefficients are used to provide speaker independence required by the objective measure. Two distortion measures are introduced for predicting speech quality based on vector quantization of the output speech record. These are the transition probability distance and the median minimum distance measure. The OBQ parameters are tested on four different speech datasets. The correlation is computed between subjective scores and the objective quality measures under a variety conditions, and the results indicate that the proposed algorithms are quite robust against speaker, text and distortion variation.

The paper investigates the effectiveness of several time alignment algorithms for application to objective speech quality estimation. The alignment issue is a common problem in objective speech quality research since proper alignment is important for obtaining accurate objective scores. Tallak et al. (1992) applied selected algorithms to synthetically distorted speech. The present paper extends these results by applying the algorithms to speech processed through the Federal Standard CELP 4800 bps coding algorithm. Simulated channel conditions introduce random and burst bit errors to further degrade speech quality. Algorithm performance is gauged by applying objective quality algorithms based on cepstral and mel-cepstral distance. Correlation coefficients between objective and subjective diagnostic acceptability measure (DAM) scores are used to determine effectiveness of the delay estimators

The effectiveness of various time alignment algorithms associated with speech is investigated. The delay estimation techniques considered here have been tested using several sets of speech data with varying types of distortion. Simple cross-correlation works best for the case of added noise, simulating a waveform coder for which it is optimal. However, for nonwaveform distortion such as T-ref simulating a low bit-rate codec, it is much less reliable. The envelope correlation seems to be consistently good for all the cases whereas the average zero crossing method seems to function more reliably for T-ref condition than the added noise case. Although, the cepstral distance method is reasonably accurate, it is computationally very intensive. Another drawback of this method is that it is not inherently a self-starting method, that is, to reduce the amount of computation with the cepstral method an approximation for the delay has to be made using one of the other methods prior to using this method. It could probably function best as a verification method. The statistical methods are less accurate than expected