| 摘要 |
A method of error concealment or correction is disclosed, suitable for signals such as speech, image and video signals, and particularly for such signals as transmitted over wireless and ATM channels. The method has improved stability for large block sizes and bunched errors. It is based on a modification of the discrete Fourier transform which is equivalent to a permutation of the Fourier coefficients. Thus, as with the conventional Fourier transform, setting a contiguous set of coefficients equal to zero in the transmitted signal enables error concealment techniques to be used at the receiver. However, with the modified transform, the zeroes are not bunched together in the spectrum, so the instability problems that arise when the conventional Fourier transform is used are mollified, and much larger block sizes can be used. <IMAGE> |
