| 主权项 |
1. A computer-implemented method performed by a data processing apparatus, the method comprising:
receiving a first version of an audio recording; receiving a second version of the audio recording; determining at least one difference between the first version of the audio recording and the second version of the audio recording using one or more of time domain analysis and frequency domain analysis; and storing the at least one difference in a difference set, wherein the difference set allows the first version of the audio recording to be distinguished from the second version of the audio recording, wherein determining at the least one difference between the first version of the audio recording and the second version of the audio recording using time domain analysis comprises:
partitioning the first version of the audio recording and the second version of the audio recording into non-overlapping blocks of fixed lengths,aligning the blocks for the second version of the audio recording with corresponding blocks for the first version of the audio recording to form block pairs,subtracting the block for the second version of the audio recording from the corresponding block for the first version of the audio recording for each block pair to obtain a residual signal,subtracting a weighted spectrum of the second version of the audio recording from the residual signal to obtain a difference signal,squaring the difference signal to obtain a squared difference signal,determining a mean value for the difference signal to obtain a threshold, andinspecting each peak in the squared difference signal that is greater than the threshold to determine if each peak represents one of the at least one differences. |
