| Development of a novel audio bandwidth extension technology with low processing power and its adoption in MPEG standards |
| Toshiyuki Nomura ・ Takeshi Norimatsu ・ Osamu Shimada |
Bandwidth extension technology with a compression rate of 1/32 existed for much lower bit rate demand. However, this bandwidth extension technology is difficult to implement in a handheld device with low computational capability, since it utilizes a complex-value filter bank consuming very high processing power. An alternative approach, the use of a real-value filter bank, can reduce the processing cost. However, it causes sound quality degradation by aliasing distortion in bandwidth extension processing. Therefore, a new high-efficiency audio compression technology, which achieves high-quality sound at a high compression rate with a low processing cost, has been in great demand for realizing the mobile broadcast service for handheld devices.
The distortion caused by the real-value filter bank is observed in all sub-bands by energy adjustment in bandwidth extension processing. However, the recipients of this Achievement Award found that the subjective sound quality degradation occurs only in specific sub-bands, and that the degree of sound quality degradation depends on the spectrum characteristic of each sub-band signal, on the basis of the knowledge obtained by the longtime research and development of various audio compression technologies. On the basis of this finding, they newly developed a detection method for sub-bands that cause sound quality deterioration by using metrics that quantify the relations of spectral characteristics between adjacent sub-bands by linear prediction. Then, the developed method reduces aliasing distortion by adaptively limiting the energy adjustment of the detected sub-bands on the basis of the quantified metrics. Finally, they achieved high-quality audio compression with 40% reduced processing cost, enabling the One-Seg broadcast service for commercial use.
The recipients of this Achievement Award have significantly contributed to standardization activities of the international standards ISO/IEC 14496-3 (MPEG) and 3GPP. Following MPEG standardization in 2003, the developed technology has been extensively utilized worldwide in various applications such as digital broadcasting, music distribution over the Internet, and digital audio players. Furthermore, the extremely high originality and effectiveness of the developed technology has resulted in its successive adoption in various audio compression standards such as the low-bit-rate compression of multichannel audio.
As described above, all the recipients of this Achievement Award performed the research and development on high-quality and low-bit-rate audio compression technology with a low processing cost and enabled the realization of a simple and convenient audio listening environment using handheld devices. Their contributions to the development of various new markets in audio broadcast, delivery, and storage through the international standardization of the developed technology are extremely remarkable and deserving of the IEICE Achievement Award.
| ・ | T. Nomura et al., “A Low-Complexity Bandwidth Extension Algorithm for MPEG-4 Audio Standardization,” General Conference of IEICE, Mar 2003. |
| ・ | O. Shimada et al., “An Aliasing Reduction Method for the MPEG-4 Audio Low Complexity Bandwidth Extension Standard,” General Conference of IEICE, Mar 2003. |
| ・ | O. Shimada et al., “A Low Power SBR Algorithm for the MPEG-4 Audio Standard and its DSP Implementation,” Preprint 6048, 116th AES Convention, May 2004. |
| ・ | O. Shimada et al., “Tradeoff Between Complexity and Memory Size in the 3GPP Enhanced aacPlus Decoder,” J. Signal Processing Syst., Vol. 57, No. 3, Dec 2009. |
