Beat Detection and Classification for Music Production Using RNN-LSTM Algorithm

Beat Detection and Classification for Music Production Using RNN-LSTM Algorithm

Devesh Kumar, Amit Kumar Thakur, Khushi Mittal, Sushmita Ray

SOET Department

K.R Mangalam University

Gurugram, India

Abstract— The process of creating music involves various techniques, including beat detection and classification. This research paper proposes a novel approach to detecting and classifying beats in music production using a recurrent neural network with long short-term memory (RNN-LSTM) algorithm. The proposed method uses a set of features that are extracted from the audio signal, including spectral flux, zero- crossing rate, and energy, to train the RNN-LSTM model. The model is trained using a large dataset of music tracks and is capable of predicting the beats and their respective classifications accurately. The proposed approach was evaluated on a dataset of various music genres, and the results showed that the RNN-LSTM algorithm outperforms traditional beat detection and classification techniques. The algorithm achieved an high accuracy as compared to the traditional techniques. Additionally, the proposed method is capable of detecting and classifying beats in real-time, making it useful for music production applications. The proposed approach has several potential applications, including music production, DJing, and music analysis. It can be used to identify the beats in a song, which can be used to synchronize different tracks, create remixes, and enhance the overall listening experience. Furthermore, the proposed method can be used to analyze the characteristics of different music genres and identify the underlying patterns that define them.

Keywords— Beat Detection, Beat Classification, Music Production, RNN-LSTM algorithm, Music Analysis.