A low-complexity and high-precision requantisation algorithm for MP3 decoding
Lin Dai, Jun Yang, Longxing Shi · International Journal of Electronics · 2009
The exponent calculation in requantisation is one of the most time-consuming operations in MP3 decoding. This article proposes a conversion method from float-point calculation to fixed-point calculation and a low-complexity and high-precision algorithm based on Newton's method for requantisation. The algorithm can be applied in many other multimedia applications to decrease the complexity. Furthermore, by means of error analysis a compensation approach is given. The experiments prove that the average error and the standard error are decreased to 0.001056% and 0.002293% respectively. The signal to noise ratio of MP3 is about 60 dB.