From Raaga to Notation: Bridging Emotions and Music Through Neural Networks

This study pioneers the integration of artificial intelligence with Indian classical music to create a neural network capable of generating raagas and musical notations based on emotional inputs. In the first phase, we trained the neural network on the ten principal thaats of Indian classical music, aiming to synthesize new raagas that maintain the traditional essence. The second phase focused on Bengali songs, using emotion indices derived from song lyrics analyzed by ChatGPT -4 to generate corresponding Swaralipi. Despite the challenges in accurately capturing the emotional depth of lyrics and translating them into musical notation, our results demonstrate the potential of neural networks in music generation and the preservation of cultural heritage. The research underscores the unexplored richness of Indian classical music and its susceptibility to being overshadowed by contemporary forms. Our findings advocate for the future use of AI in safeguarding and revitalizing the legacy of Indian classical music traditions.