| 摘要 |
<p>A method of echo cancellation, particularly for use in high definition applications, splits an input signal and reference signal into M single-sided sub-band. The subbanded signals are downsampled at a downsampling rate N, where N@M, adaptively filtered, and recombined to produce an output signal. The sub-bands are preferably oversampled such that N<M. The use of oversampling and single-sided sub-banding reduces complexity and avoids aliasing problems.</p> |
