摘要 |
<p>Cross talk terms may be removed from speech signals in a multi-party conference system having a plurality of users. In a preferred embodiment of the present invention, an estimator (210, 310) is used to first estimate the respective cross talk terms for each of the said plurality of users. These terms are then multiplied with a filter (230, 330) based on the estimated cross talk terms. Source speech signals are thereby recovered from the plurality of multi-party conference users without the cross talk.</p> |