Forensic Synthetic Speech Inspection Technique based on Non-formant speech features

With the development of speech synthesis technology, the simulation of specific individual’s speech has gradually matured, synthetic speech is easily and perceptually recognized as real speech, which may occur frequently in illegal activities. At present, the synthetic speech synthesis speech detection method does not fully take into account the complexity of the actual application, and has not yet formed a systematic and effective detection means. To this end, two young males and two young females from different dialect regions were recruited to read the same text. Their voices were recorded and used by the speech synthesis software to generate four synthetic voices. This paper compares the non-formant speech characteristics of synthesized speech with those of natural speech in terms of verbal speech features, verbal speech and articulation defect features, and sound source features. The results of the experiment show that the above non-formant speech features could efficiently detect the synthetic speech from the real speech.