Architecture design of MDCT-based psychoacoustic model co-processor in MPEG advanced audio coding

A novel architecture of MDCT-based Psychoacoustic Model (PAM) in an MPEG-2/4 Advanced Audio Coding (AAC) encoder is proposed. PAM, a very important role in AAC encoding, not only determines sound quality of a given encoder but also influences a lot in computational complexity. Our paper presents a dedicated hardware design, which is different from traditional programmable approaches. Moreover, to reduce the complexity, we replace the calculations of spreading function with reduced fixed coefficients, and decrease the transform kernels from three to two units.