Judicial expertise of speaker identity based on improved pitch algorithm
Zhijun Li, Hu Xuelong, Guo Na · 2017
Based on the comprehensive analysis and comparison in the characteristic of acoustics between the sample speaker and the case sample speaker, we can get the identification of identity. The identification can be used in many cases like providing clues, directions and ranges to investigate and solve a case. In addition, it can also provide independent evidence for litigation. As one of the main features of speaker, fundamental frequency curve plays an important role in the identification of identity. In this passage, we create an improved pitch period extraction algorithm based on the traditional way of detection method of pitch period. This experiment has proved that the method in this passage works more accurately than the traditional way in the noise circumstance and it can provide more convincing evidence for judicial expertise.