| 摘要 |
Automatic detection of errors among different formatted sound tracks of the same language on a motion picture film stock can be achieved by first acquiring successive audio segments from each of the sound tracks. During a time window of prescribed duration, the audio of each different formatted track undergoes analysis to yield a numeric value. The successive analysis of the audio continues until no further audio exists for analysis. The resultant collection of numeric values undergoes formatting into a numeric file for comparison against a reference file representing audio obtained from a particular source, such as originally recorded material, a sound print, or a duplicated copy of a sound film. If the difference between a formatted numerical file and the reference file exceeds a threshold value, then an error exists in that formatted sound track, and an operator can take appropriate action. |
