Soo-Ngee Koh

Subband coding of high-fidelity quality audio signals at 128 kbps

The design and performance of a 24-band subband coder for the encoding of high-fidelity quality 20 kHz audio signals at 128 kb/s are considered. The slightly different performances of the subband coder using two different adaptive bit allocation algorithms, namely the modified Ramstad algorithm and the algorithm based on noise masking threshold, are discussed and compared in terms of segmental signal-to-noise ratio, mask-to-noise ratio, and the probability of noise measures.<<ETX>>

A 5.4 kbps speech coder based on multi-band excitation and linear predictive coding

Recent work has shown that multi-band excitation (MBE) is capable of synthesizing high quality speech in the range of 4.8-8.0 kbps. In this paper, the multi-band excitation and linear predictive coding (MBE-LPC) model, which used the LPC analysis to obtain the prediction residual, and uses the MBE model to estimate the residual spectrum is presented. Our motivation is to improve the excitation model of LPC vocoders in the frequency domain. Based on the MBE-LPC model, a 5.4 kbps speech coding is presented. An adaptive postfilter is used to improve the perceptual quality of the decoded speech. Informal listening tests show that the perceptual quality of the decoded speech of the proposed coder is better than that of the 4.15 kbps improved (IMBE) coder.<<ETX>>

A neural network based perceptual audio coder

The implementations and performance results of a neural network based perceptual audio coder is reported. The coder uses the configuration of the ISOIMPEG audio layer II coder with the perceptual analysis block replaced by a 2 layer network trained to estimate the masking thresholds required for the bit allocation. The 2 layer network is trained by a back propagation algorithm using the energies in the subbands as inputs and the masking thresholds of obtained from psychoacoustic model II of the ISOIMPEG audio coder as the reference outputs. The result is a coder which performs favourably in quality against the ISOIMPEG audio layer 2 coder at bit rates of 256 kbit/s and 192 kbit/s stereo. Performance at a bit rate of 128 kbit/s stereo was however, found to be poorer.<<ETX>>

Phase unwrapping methods for quadratic phase interpolation in voiced speech sythesis

This paper considers the use of a quadratic polynomial for phase interpolation in voiced speech synthesis using the multi-band excitation (MBE) model. The problem of phase unwrapping when using quadratic interpolation, originally proposed by Griffin and Lim (1988), is examined. In addition to that, another three new methods for phase unwrapping are studied. For the purpose of comparing the different unwrapping methods used for quadratic interpolation, the four unwrapping methods are used to synthesize voiced speech using the MBE model. Both the theoretical analysis and simulation results show that the fourth phase unwrapping method outperforms the other methods for voiced speech synthesis.