| 摘要 |
A high efficiency encoding audio signals is disclosed. The system includes a signal energy calculating circuit for calculating allowable noise energies by utilizing characteristics of human acoustic sense and calculating allowable noise energies of a block under consideration by weighted addition of the allowable noise energies calculated for at least one past block with respect to the block under consideration, a first-order allowable noise energy calculating unit, a second-order allowable noise energy calculating unit, a calculating circuit for calculating characteristics of human acoustic sense, and a memory. Adaptive quantization is performed in the quantization circuit with the bit allocation number which is based on the calculated allowable noise energies. Signals undergoing large temporal fluctuations can be encoded efficiently, while aurally outstanding noise may be reduced to improve the sound quality by eliminating sound quality deterioration known as pre-echo.
|
