Please refer to this paper if you are using this TF-representation:
(In revision) M. Nigro, A. Rueda and S. Krishnan, "Acoustic Scene Classification Using Time–FrequencyEnergy Emphasis and Convolutional Recurrent NeuralNetworks", Advs in Intelligent Syst., Computing, Vol. 1361, S. Chandramohan et al. (Eds): Artificial Intelligence and Evolutionary Computations in Engineering Systems, 978-981-16-2673-9, 494986_1_En, (Chapter 21)