The FIR-based reverberators, which convolve the input sequence with an impulse response modelling the concert hall, have better quality compared to the IIR-based approach. However, the high computational complexity of the FIR-based reverberators limits the applicability to most cost-oriented system. This paper introduces a method that uses perceptual criterion to reduce the complexity of convolution methods for reverberation. Also, an objective measurement criterion is introduced to check the perceptual difference from the reduction. The result has shown that the length of impulse response can be cut off by 60% without affecting the perceptual reverberation quality. The method is well integrated into the existing FFT-based approach is have around 30% speed-up.

Download

This paper reports on a new multi-label classification task for guitar effect recognition that is closer to the actual use case of guitar effect pedals. To generate the dataset, we used multiple clean guitar audio datasets and applied various combinations of 13 commonly used guitar effects. We compared four neural network structures: a simple Multi-Layer Perceptron as a baseline, ResNet models, a CRNN model, and a sample-level CNN model. The ResNet models achieved the best performance in terms of accuracy and robustness under various setups (with or without clean audio, seen or unseen dataset), with a micro F1 of 0.876 and Macro F1 of 0.906 in the hardest setup. An ablation study on the ResNet models further indicates the necessary model complexity for the task.

Download