Perceptual Neural Physical Sound Matching

Description

This website demonstrates the sound matching result of EfficientNet-B0 type models trained by various loss functions. Each row displays five sound matching outputs of the same target sound (leftmost).

Sound Matching outputs

FTM

Target Current Ploss PNP ICASSP DDSP Masuda and Saito

AM/FM arpeggiator

Target Current Ploss PNP ICASSP DDSP Masuda and Saito