Assessment of Controller Speech Quality Based on Deep Learning
With the rapid development of communication technology,the functions and scale of voice communi-cation systems are also undergoing rapid changes.At the same time,various encoding and decoding technologies are emerging one after another.Different types of encoding and decoding technologies and transmission methods can cause varying degrees of damage to speech quality.The quality of voice during communication between controllers and pilots directly affects the efficiency of communication between both parties and the safety of aircraft operation.To ensure the quality of voice signals during communication,this article considers real-time evaluation of voice signals.However,subjective evaluation methods have high costs and significant subjective factors.Therefore,it is considered to select objective speech quality evaluation methods for evaluating controlled speech.Deep learning based speech quality assessment is a reference free speech quality assessment method that is precisely suitable for objective evaluation of directly recorded controlled speech.
speech quality evaluationdeep learningneural networksair traffic control voice
万方数据
维普
