Automatic Visual Effect Adjustment System

This paper presents a novel approach to enhancing the audience experience at live concerts by synchronizing visual effects with the music. We introduce an automatic visual effect adjustment system designed and implemented for real-life applications. Our method leverages wavelet transform in conjunction with a microcontroller unit (MCU) to capture and analyze the rhythm of live concert music. The visual effects, represented by LED strips, are dynamically modulated based on the extracted rhythm. However, suboptimal settings during wavelet transform may introduce rhythm distortion. To address this issue, we incorporate a music genre recognition system into our embedded setup to fine-tune the wavelet transform parameters. This recognition system converts audio signals into log-power spectrograms and employs a deep learning (DL) model to classify music genres. The output of the genre recognition system is used to optimize the rhythm extraction settings, ensuring optimal synchronization between the visual effects and the live concert music. The identified genre and extracted rhythm jointly drive the behavior of the LED strips, creating a visually immersive experience for the audience.