A Deep Learning Loss Function Based on the Perceptual Evaluation of the Speech Quality Martín Doñas, Juan M. Gómez García, Ángel Manuel González López, José Andrés Peinado Herreros, Antonio Miguel Deep learning (DL) Speech enhancement This letter proposes a perceptual metric for speech quality evaluation, which is suitable, as a loss function, for training deep learning methods. This metric, derived from the perceptual evaluation of the speech quality algorithm, is computed in a per-frame basis and from the power spectra of the reference and processed speech signal. Thus, two disturbance terms, which account for distortion once auditory masking and threshold effects are factored in, amend the mean square error (MSE) loss function by introducing perceptual criteria based on human psychoacoustics. The proposed loss function is evaluated for noisy speech enhancement with deep neural networks. Experimental results show that our metric achieves significant gains in speech quality (evaluated using an objective metric and a listening test) when compared to using MSE or other perceptual-based loss functions from the literature. 2021-11-15T07:28:13Z 2021-11-15T07:28:13Z 2018-09-19 info:eu-repo/semantics/article Martín-Doñas, J. M., Gomez, A. M., Gonzalez, J. A., & Peinado, A. M. (2018). A deep learning loss function based on the perceptual evaluation of the speech quality. IEEE Signal processing letters, 25(11), 1680-1684. http://hdl.handle.net/10481/71497 10.1109/LSP.2018.2871419 eng http://creativecommons.org/licenses/by-nc-nd/3.0/es/ info:eu-repo/semantics/openAccess Atribución-NoComercial-SinDerivadas 3.0 España IEEE