A General Compression Approach to Multi-Channel Three-Dimensional Audio

Abstract

, "A general compression approach to multi-channel three-dimensional audio," IEEE Transactions on Audio, Speech and Language Processing, vol. 21, (8) pp. 1676-1688, 2013 A general compression approach to multi-channel three-dimensional audio AbstractThis paper presents a technique for low bit rate compression of three-dimensional (3D) audio produced by multiple loudspeaker channels. The approach is based on the time-frequency analysis of the localization of spatial sound sources within the 3D space as rendered by a multi-channel audio signal (in this case 16 channels). This analysis results in the derivation of a stereo downmix signal representing the original 16 channels. Alternatively, a mono-downmix signal with side information representing the location of sound sources within the 3D spatial scene can also be derived. The resulting downmix signals are then compressed with a traditional audio coder, resulting in a representation of the 3D soundfield at bit rates comparable with existing stereo audio coders while maintaining the perceptual quality produced from separate encoding of each channel. © 2006-2012 IEEE.

JND-based spatial parameter quantization of multichannel audio signals

“…The inter-channel spatial parameters such as ICLD can be extracted from only two channel signals. Cheng proposed a method to extract the azimuth of virtual sound from an arbitrary number of loudspeaker signals [22]. In Cheng's method, downmix signal of multichannel signals were obtained as the signal of virtual sound source.…”

“…Cheng designed uniform quantization codebooks with fixed step sizes of 2°or 3°for azimuths in horizontal plane. Since high quantization resolution of azimuths in Cheng's method, the quantization codebook sizes for azimuths in a full circle were almost twice or triple of that in MPEG surround and resulted in high bit rates of spatial parameters [22].…”

“…2 [22]. Given that only one virtual sound source generated by loudspeakers exists for each time-frequency bin, multichannel signals can be represented with a virtual sound source with its spatial location information.…”

“…Except ICLD, the azimuth of virtual sound is also commonly used as one of the spatial parameters as in [22], especially used for more than two loudspeakers. The azimuth of virtual sound extracted from two loudspeaker signals by:…”

“…3 Quantization codewords distributions in different methods. a High resolution quantization of spatial parameters in MPEG Surround [15], denoted as MPEG_H; b Low resolution quantization of spatial parameters in MPEG Surround [15], denoted as MPEG_L; c Choi's method [19]; d High resolution quantization in Cheng's method [20], denoted as Cheng_H; e Low resolution quantization in Cheng's method [20], denoted as Cheng_L; f Low resolution quantization in SLQP [22], denoted as SLQP_L characteristics of different azimuths in a full circle. These quantization methods of spatial parameters will result in either spatial perceptual distortions of virtual sound images, or inefficient compressions of spatial parameters.…”