| 摘要 |
<p>In a method and apparatus for interference suppression in speech signals for speech recognition purposes, the speech being converted along with other sound phenomena into an electric signal (summation signal) and the other sound phenomena consisting of noises and a sound phenomenon produced by electro-acoustic conversion of an audio signal, the noises are furthermore recorded and converted into a noise signal. The noise signal and the audio signal are passed via adaptive filters and subtracted from the summation signal. The resulting signal is compared with the noise signal on the one hand and with the audio signal on the other hand. The results of the comparison are used to control the adaptive filters. <IMAGE></p> |
