FM Broadcast Audio Signal Analysis Using Time–Frequency Distribution Function

Abstract

For any product, the end customer is at highest priority with regard to available user options. Presently, commercial radio broadcast contains more number of advertisements instead of content. The paper presents a solution to detect advertisement in between duration for which songs being played. With the recent advancements in the area of audio processing and analysis, the analysis of songs and advertisements is made by studying real-time data collected from various radio broadcasts. After examining each particular characteristics of decomposed signal in time, frequency, time–frequency, and statistical domains for different audio clippings, it is observed that there is a pattern between the songs and advertisements. This data can be used for the classification of songs and advertisements. After correct classification, a software-based routine can be made to automatically switch to another station in case of advertisement being played.