摘要 |
The present invention relates to the improvement of sound quality. Provided is a method and an apparatus for determining a probabilistic sound period for speech discrimination in a multiple-microphone-based sound quality enhancement system. According to the present invention, the method can provide a constant tone by reacting adaptively even when the type of noise is changed and have a small loss of voice in an unusual noise environment of a low signal to noise ratio (SNR). Additionally, the present invention can correspond to impulse noise in real time and can process the noise in real time even if the type of the noise is abruptly changed. Also, the present invention uses low power according to the small amount of operations, provides a solution as a software type library to increase versatility, and requires tuning in a short time. |