Audio processing method and device, and training method and device of audio processing model

• Main Authors: ZHANG CHEN, ZHENG XIGUANG, XING WENHAO
• Format: Patent
• Language: Chinese; English
• Published: 11.10.2022
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS

The invention relates to an audio processing method and device and a training method and device of an audio processing model, and the audio processing method comprises the steps: obtaining network service quality features in a process of receiving a plurality of audio packets of target audio; inputting the network service quality features into an audio processing model to obtain a recommendation length; wherein the audio processing model is obtained by training according to audio samples, network service quality characteristics obtained in the audio sample receiving process and received audio obtained by carrying out jitter buffering processing on the audio samples; adjusting a jitter buffer area according to the recommended length; and performing jitter buffer processing on the received audio packet according to the adjusted jitter buffer area. 本公开关于一种音频处理方法和装置、音频处理模型的训练方法和装置，音频处理方法包括：在接收目标音频的多个音频包的过程中，获取网络服务质量特征；将网络服务质量特征输入音频处理模型，得到推荐长度；其中，音频处理模型是根据音频样本、接收音频样本过程中获取的网络服务质量特征、音频样本经抖动缓冲处理得到的接收音频训练得到的；根据推荐长度，调整