Deep Learning applied to Speech Transmission Index Prediction: Simulations and Measurements

This is the database that was used to evaluate acoustic quality in 5 classrooms of the Federal University of Paraná (UFPR) at the Polytechnic Campus, applying the standards ISO 9921, ISO 3382, NBR 12179, NBR 10152, and IEC 602068-16. Thus, it was determined the possible correlations and redundancies between the following descriptors: Reverberation Time - T30, Central Time - Ts, Early Time Decay - EDT, Definition - D50, Clarity - C50, Useful-to-detrimental sound ratio - U50 and Speech Transmission Index - STI, and the significance of the following factors on those: background noise - (A), sound absorption coefficient - (B), confinement - (C) and occupation - (D). The methodology consisted of applying the Design of Experiments of type 24 not replicated, based on the validated simulations in the ODEON version 11 software, to create the response matrices, totaling 80 virtual rooms and 53 responses. A correlation matrix between the descriptors was created and then the Principal Components Analysis was applied. We are pleased to submit the data proof document that accompanies our manuscript titled “ Combined evaluation of room acoustic descriptors in different structural configurations via ODEON simulations and Artificial Neural Networks “which we have submitted for consideration in the Archives of Acoustics journal. The purpose of this data proof is to provide detailed evidence supporting the results and conclusions presented in our manuscript.