The USTC System for Cadenza 2024 Challenge

Hongbo Lan, Tianyou Cheng, Maokui He, Hang Chen, Jun Du

Abstract: This paper reports our submission to the ICASSP 2024 Cadenza Challenge, focusing on the non-causal system. The challenge aims to develop a signal processing system that enables personalized re-balancing of music to improve the listening experience for individuals with hearing loss when they listen to music via their hearing aids. The system is based on the Hybrid Demucs model. We fine-tuned the baseline model on the given dataset with a multi-target strategy and added a mixture of information in the downmixing stage. We combined different models and achieved a score of 0.6878 in the Hearing Aid Audio Quality Index (HAAQI) metric on the validation set and 0.5929 on the test set.