2021/03/02 by Junghyun Koo, Koo, Junghyun, Seungryeol Paik +3 · 5 citations
Computer Science · #Audio and Speech Processing (eess.AS) #FOS: Computer and information sciences #FOS: Electrical engineering #Machine Learning (cs.LG) #Music Technology and Sound Studies #Music and Audio Processing #Sound (cs.SD) #Speech and Audio Processing #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2103.02147
openalex publication_date 2021/03/02 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Reverb plays a critical role in music production, where it provides listeners\nwith spatial realization, timbre, and texture of the music. Yet, it is\nchallenging to reproduce the musical reverb of a reference music track even by\nskilled engineers. In response, we propose an end-to-end system capable of\nswitching the musical reverb factor of two different mixed vocal tracks. This\nmethod enables us to apply the reverb of the reference track to the source\ntrack to which the effect is desired. Further, our model can perform\nde-reverberation when the reference track is used as a dry vocal source. The\nproposed model is trained in combination with an adversarial objective, which\nmakes it possible to handle high-resolution audio samples. The perceptual\nevaluation confirmed that the proposed model can convert the reverb factor with\nthe preferred rate of 64.8%. To the best of our knowledge, this is the first\nattempt to apply deep neural networks to converting music reverb of vocal\ntracks.\n