CPC G10L 25/18 (2013.01) [G10L 25/30 (2013.01)] | 12 Claims |
1. A method of processing audio data, the method comprising:
inputting spectral data of the audio data into a feature extraction model to obtain a first feature vector;
inputting the first feature vector into a periodic/aperiodic indicator detection model to obtain a periodic/aperiodic indicator, wherein the periodic/aperiodic indicator is a feature vector indicating periodic audio data of the first feature vector and aperiodic audio data of the first feature vector;
concatenating the first feature vector and the periodic/aperiodic indicator to obtain a second feature vector;
inputting the second feature vector into a fundamental frequency detection model to obtain a fundamental frequency;
concatenating the first feature vector, the periodic/aperiodic indicator and the fundamental frequency to obtain a third feature vector;
obtaining a spectral energy according to the third feature vector;
obtaining a harmonic structure of the audio data according to the fundamental frequency and the spectral energy;
obtaining a noise information in the audio data according to the first feature vector; and
obtaining synthetic audio data according to the harmonic structure and the noise information.
|