Research on Voice Humming to Guzheng Audio Transcription and Pitch-Lowering Processing Based on Genetic Algorithm and YIN Algorithm

In recent years, AI-based music generation and transformation technologies have received widespread attention. This study presents an automated system for transcribing vocal humming audio into Guzheng audio, integrating the YIN algorithm and genetic algorithm, and optimizing pitch and timbre matching through pitch-lowering processing. First, the YIN algorithm is used to extract the fundamental frequency of the humming audio, and then the genetic algorithm optimizes the frequency sequence to better align with the Guzheng's pitch range. To further improve the audio generation quality, the system introduces pitch-lowering processing, adjusting the generated audio frequencies to fit within the Guzheng's pitch range. Experimental results show that pitch-lowering processing significantly enhances the accuracy of the pitch and timbre matching. Moreover, the genetic algorithm outperforms the YIN algorithm in optimizing pitch and timbre, with the generated audio exhibiting a more realistic Guzheng timbre. Ultimately, the system achieves significant improvements in pitch, timbre, and auditory effects. The proposed transcription scheme demonstrates good applicability in musical applications involving pitch range conversion and provides new research insights for automated instrument transcription.