Dataset of sound field simulations above finite absorbers

The authors' documentation on the training, validation, and testing datasets in their paper "<em>Sound absorption estimation of finite porous samples with deep residual learning</em>." Goal: predict sound absorption coefficients from array measurements above finite porous materials. In effect, the model learns relevant features from edge-diffracted sound fields to predict absorption as if the sample were infinite. Results: Using the present datasets, the method is shown to perform at least as well as the two-microphone method in the higher frequency range (weaker contribution of edge diffraction to the sound field) More details on the models, and reproduction of results can be found in the GitHub repo: https://github.com/eliaszea/finite-absorber-ML.