2018/09/14 by Leonardo Gabrielli, Gabrielli, Leonardo, Stefano Tomassetti +7 · 1 citation
Computer Science · #Audio and Speech Processing (eess.AS) #FOS: Computer and information sciences #FOS: Electrical engineering #Machine Learning (cs.LG) #Machine Learning (stat.ML) #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.1809.05483
openalex publication_date 2018/09/14 · openalex created_date 2022/08/03 · openalex updated_date 2026/07/28
One of the challenges in computational acoustics is the identification of\nmodels that can simulate and predict the physical behavior of a system\ngenerating an acoustic signal. Whenever such models are used for commercial\napplications an additional constraint is the time-to-market, making automation\nof the sound design process desirable. In previous works, a computational sound\ndesign approach has been proposed for the parameter estimation problem\ninvolving timbre matching by deep learning, which was applied to the synthesis\nof pipe organ tones. In this work we refine previous results by introducing the\nformer approach in a multi-stage algorithm that also adds heuristics and a\nstochastic optimization method operating on objective cost functions based on\npsychoacoustics. The optimization method shows to be able to refine the first\nestimate given by the deep learning approach and substantially improve the\nobjective metrics, with the additional benefit of reducing the sound design\nprocess time. Subjective listening tests are also conducted to gather\nadditional insights on the results.\n