Automatic quality control of broadcast audio

Abstract

For many years the broadcast industry has been relying on limited and manual quality control (QC) processes for audio. Unlike for video where some automated solutions exist there has not been much work in automated audio QC processes. Any audio functionality that does exist in audio QC is typically limited to basic audio attributes such as loudness or audio silence detection with very little attention given to automated analysis of perceptual aspects such as audio anomalies or speech intelligibility. This paper describes work undertaken as part of the AQUA project funded by InnovateUK to address this with an automated software solution for both production and distribution of audio content on premises or in the cloud. Of particular interest to this paper is the analysis of speech content for intelligibility and expletive detection and audio anomaly detection and compliance across multiple audio formats.