Artificially Synthesising Data for Audio Classification and Segmentation to Improve Speech and Music Detection in Radio Broadcast

Date
2021-05-07Author
Subject
Metadata
Show full item recordAbstract
Segmenting audio into homogeneous sections such as music and speech helps us understand the content of audio. It is useful as a pre-processing step to index, store, and modify audio recordings, radio broadcasts and TV programmes. Deep learning models for segmentation are generally trained on copyrighted material, which cannot be shared. Annotating these datasets is time-consuming and expensive and therefore, it significantly slows down research progress. In this study, we present a novel procedure that artificially synthesises data that resembles radio signals. We replicate the workflow of a radio DJ in mixing audio and investigate parameters like fade curves and audio ducking. We trained a Convolutional Recurrent Neural Network (CRNN) on this synthesised data and outperformed state-of-the-art algorithms for music-speech detection. This paper demonstrates the data synthesis procedure as a highly effective technique to generate large training sets for deep neural networks.
Collections
Publisher
Journal
Volume
Pagination
Conference name
Start date
Finish date
Publisher URL
Recommended, similar items
The following license files are associated with this item: