| 摘要 |
An echo canceller includes echo signal canceling sections, first level detection sections, second level detection sections, third level detection sections, and priority decision section. The echo signal canceling sections are independently provided to echo paths of speech communication paths to learn and update the estimated values of impulse responses on the echo paths and cancel echoes generated in the corresponding speech communication paths on the basis of the updated estimated values of the impulse responses. The first level detection sections detect the levels of transmission speech signals. The second level detection sections detect the levels of residual signals obtained by removing echo components from the transmission speech signals. The third level detection sections detect the levels of reception speech signals. The priority decision section decides the priorities of the speech communication paths on the basis of a detection output from at least one of the first to third level detection sections and selects the echo signal canceling sections in accordance with the decided priorities, thereby enabling a learning operation of the estimated values of the impulse responses on the echo paths.
|
