Sunday, October 6 2:00 pm 4:30 pm
SESSION H: LOW BIT-RATE CODING, PART 1
Chair: Gary Brown, Tensilica, Santa Clara, CA, USA
H-1 Scalable Lossless Audio Coding Based on MPEG-4 BSACDoh-Hyung Kim, Jung-Hoe Kim, Sang-Wook Kim, Samsung Advanced Institute of Technology, Suwon, Korea
In this paper a new hybrid type of scalable lossless audio coding scheme based on MPEG-4 BSAC (bit sliced arithmetic coding) is proposed. This method introduces two residual error signals, lossy coding error signal and prediction error signal, and utilizes the rice coding as a lossless coding tool. These kinds of processes enable an increase in the compression ratio. As a result of experiment, average total file size can be reduced about 50 to 60 percent of the original size. Consequently, a slight modification of the conventional MPEG-4 general audio coding scheme can give a scalable lossless audio coding functionality between lossy and lossless bitstream.
Convention Paper 5679
H-2 Lossless Audio Coding Using Adaptive Multichannel PredictionTilman Liebchen, Technical University of Berlin, Berlin, Germany
Lossless audio coding enables the compression of digital audio data without any loss in quality due to a perfect reconstruction of the original signal. The compression is achieved by means of decorrelation methods such as linear prediction. However, since audio signals usually consist of at least two channels, which are often highly correlated with each other, it is worthwhile to make use of inter-channel correlations as well. This paper shows how conventional (mono) prediction can be extended to stereo and multichannel prediction in order to improve compression efficiency. Results for stereo and multichannel recordings are given.
Convention Paper 5680
H-3 Design of Error-Resilient Layered Audio CodecDai Yang, Hongmei Ai, Chris Kyriakakis, C.-C. Jay Kuo, University of Southern California, Los Angeles, CA, USA
Current high quality audio coding techniques mainly focus on coding efficiency, which makes them extremely sensitive to channel noise, especially in high error rate wireless channels. In this paper we propose an error-resilient layered audio codec (ERLAC) which provides functionalities of both fine-grain scalability and error-resiliency. A progressive quantization, a dynamic segmentation scheme, a frequency interleaving technique, and an unequal error protection scheme are adopted in the proposed algorithm to construct the final error robust layered audio bitstream. The performance of the proposed algorithm is tested under different error patterns of WCDMA channels with several test audio materials. Our experimental results show that the proposed approach achieves excellent error resilience at a regular user bit rate of 64 kb/s.
Convention Paper 5681
H-4 Application of a Concatenated Coding System with Convolutional Codes and Reed-Solomon Codes to MPEG Advanced Audio Coding Dong Yan Huang1, Say Wei Foo2, Weisi Lin3, Ju-Nia Al Lee4 - 1Institute of Microelectronics, Singapore, Singapore; 2Nanyang Technological University, Singapore, Singapore; 3Laboratories of Information Technology, Singapore, Singapore; 4National University of Singapore, Singapore
Reliable delivery of audio bitstream is vital to ensure the acceptable audio quality perceived by 3G network customers. In other words, an audio coding scheme that is employed must be fairly robust over the error-prone channels. Various error-resilience techniques can be utilized for the purpose. Due to the fact that some parts of the audio bitstream are less sensitive to transmission errors than others, the unequal error protection (UEP) is used to reduce the redundancy introduced by error resilience requirements. The current UEP scheme with convolutional codes and multistage interleaving has an unfortunate tendency to generate burst errors at the decoder output as the noise level is increased. A concatenated system combining Reed-Solomon codes with convolutional codes in the UEP scheme is investigated for MPEG advanced audio coding (AAC). Under severe channel conditions with random bit error rates of up to 5x10-02, the proposed scheme achieved more than 50 percent improvement in residual bit error rate over the original scheme at a bit rate of 64 kb/s and sampling frequency of 48 kHz. Under burst error conditions with burst error length of up to 4 ms, the proposed scheme achieved more than 65 percent improvement in bit error rate over the original scheme. The average percentage overhead incurred by using the concatenated system is about 3.5 percent of the original UEP scheme. Further improvements are made by decreasing the puncturing rate of convolutional codes. However, this can only be adopted when high protection is needed in extremely noisy conditions (e.g., channel BER significantly exceeds 1.00e-02) since it incurs increased overheads.
Convention Paper 5682
H-5 A Simple Method for Reproducing High Frequency Components at Low Bit-Rate Audio CodingJeongil Seo, Daeyoung Jang, Jinwoo Hong, Kyeoungok Kang, Electronics & Telecommunications Research Institute (ETRI), Yuseong-Gu, Deajon, Korea
In this paper we describe a simple method for reproducing high frequency components at low bit-rate audio coding. To compress an audio signal at low bit rates (below 16-kb/s per channel) we can use a lower sampling frequency (below 16 kHz) or high performance audio coding technology. When an audio signal is sampled at a low frequency and coded at a low bit rate, high frequency components and reverberant sound are lost because of quantization noise between pitch pulses. In a short-term period, the harmonic characteristic of audio signals is stationary, so the replication of high-frequency bands with low-frequency bands can extend the frequency range of resulting sound and enhance the sound quality. In addition, for reducing the number of bands to be reproduced we adapted this algorithm at the Bark scale domain. For compatibility with a conventional audio decoder, the additional bitstream is added at the end of each frame, which is generated by a conventional audio coder. We adapted this proposed algorithm to MPEG-2 AAC and increased the quality of audio in comparison with the conventional MPEG-2 AAC coded audio at the same rate. The computational cost of the proposed algorithm is similar to or a little more than a conventional MPEG-2 AAC decoder.
Convention Paper 5683