Voice scoring method and voice scoring device

• Main Authors: XIE RUI, CHEN XIXING, ZENG RUIHONG, XIONG JIA, LI ZIJUN, JIAO NANKAI, CHEN GUANGYAO, PAN ZISHENG, MA JINLONG, HUANG XIANGKANG, XU ZHIJIAN
• Format: Patent
• Language: Chinese; English
• Published: 18.04.2023
• Subjects: ACOUSTICS; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

According to the voice scoring method and the voice scoring device provided by the invention, the to-be-detected audio is input into the quality detection model, and the audio quality is evaluated by using a traditional detection algorithm, so that the scoring processing time is saved, the to-be-detected audio can be input into the activity detection model only when the audio quality meets a certain condition, and the scoring efficiency is improved. The voice detection algorithm based on deep learning is utilized to detect the voice activity so as to ensure the accuracy of voice scoring, only when the voice activity meets certain conditions, the to-be-detected audio can be input into the pronunciation detection model, and the voice detection algorithm based on the neural network is utilized to detect the accuracy of voice pronunciation. Therefore, the score of the to-be-tested audio is determined according to the voice activity and the voice pronunciation accuracy, so that the voice score accuracy is further