Estimation of the road traffic sound levels based on Non-Negative Matrix Factorization dataset

Sound database to compute the NonnegMatrixFact experience (download the .m files here: https://github.com/jean-remyGloaguen/article2017EstimationAmbiance) dedicated to the estimation of the traffic sound level on urban sound mixtures.

This sound database includes two subfolders: ambiance and dictionary

- ambiance folder includes 6  sub-corpus (alert, animals, climate, human, mechanics and transportation). In each sub-corpus, 50 sound mixtures are available and divided into two files: one dedicated to the traffic signal, the other to the interfering signal.

- dictionary folder includes the isolated sounds dedicated to the dictionary learning for NMF