Vectorial Spectral Quantization for Audio Coding

Abstract

The paper introduces a new coding methodology of the spectral modified discrete cosine transform (MDCT) coefficients of an audio signal. A lattice quantizer is used for each spectral sub-band, having the dimension equal to the size of the respective sub-band. The information that needs to be encoded consists of lattice codevector indexes, side information relative to the number the bits on which the indexes are represented and the integer exponents of the sub-band scaling factors. The nature of the side information, together with the parameterization of the quantization resolution allows the use of the method for a large range of bitrates e.g. for 44.1kHz sampled mono files, from 128 kbits/s down to 16 kbits/s. Subjective listening tests show similar performance of the proposed method to the advanced audio coding (AAC) codec for high bitrates (128 kbits down to 64 kbits/s) and clearly better performance for lower bitrates.

New Directions in Lattice Based Lossy Compression

“…For the sake of completeness, we present briefly the overall audio encoding framework that uses rectangular lattice truncations for quantization. For a detailed description see (Vasilache & Toukomaa, 2006). The overall performance of the audio coding method is similar to the MPEG4-AAC for higher bitrates (128kbits/s down to 64kbits/s) and better than MPEG4-AAC for lower bitrates.…”

“…The number of bits for the quantized spectral coefficients are calculated according to the formulas from Equations 11 and 13. (Vasilache & Toukomaa, 2006) Compared with the number of bits per frame available for spectral quantization only, in the fixed rate case, for the considered bitrates, the values in Table 3 give an average of 30% bitrate reduction without any loss of quality. The method (labeled as 'LatVQ') without entropy coding was compared in (Vasilache & Toukomaa, 2006), against the quantization procedure from the MPEG4-AAC codec, in a MUlti Stimulus test with Hidden Reference and Anchor (MUSHRA) (BS.1534(BS.…”

“…We have considered two encoding bit rates 32kbits/s and 48kbits/s for the audio codec from (Vasilache & Toukomaa, 2006). The number of bits for the quantized spectral coefficients is calculated according to the formulas from Equations 11 and 12.…”

“…Audio coding methods based on lattice quantization have been as well proposed (Vasilache & Toukomaa, 2006).…”