A multi-resolution spectrogram approach for estimating the physical parameters of a plate reverb
En palabras de los autores
The ResNet-18 image classification model is employed to determine the physical parameters of a plate reverb from a recording of the impulse response. The model is adapted to derive parameters using normalized and downsampled multi-resolution spectrograms computed from the provided impulse responses (IRs). To refine the prediction of the output location, the spectral phase response is also included as an additional input channel to the network since multiple output locations can give the same magnitude response for high-order resonant modes. On a 5000 IR validation set, our model achieves an average normalized mean squared error (NMSE) of 0.02920 across all parameters, with the lowest average NMSE occurring for parameters yo (0.00228), Ly (0.00347), and xo (0.00574).
Apareció: jueves, 24 de septiembre. arXiv. Preprint con versión publicada.
Versión publicada: Proc. 29th Int. Conf. Digital Audio Effects (DAFx26), 2026
Comentario de los autores: Parameter Estimation Challenge, Task A