PIANO-LAB: A SOFTWARE PACKAGE FOR SIGNAL PROCESSING OF SYSTEMATICALLY COLLECTED GRAND PIANO VIBROACOUSTIC DATA

Quality assessment is a crucial step when making musical instruments. The uncertainties and the subjectivity typical of perception-based testing need to be overcome in an industrial context, where production quality calls for a systematic and reliable way to support the design of products with stable and ever higher performance standards. In the framework of their long-term research cooperation, Politecnico di Milano and Fazioli Pianoforti have developed a modular software package for the processing, analysis and visualization of grand piano vibroacoustic data. The software is intended to query the database of the experimental data that are being systematically collected through a dedicated measurement procedure, each time a grand piano reaches the end of the production process. The use of objective vibroacoustic data allows quantitative and physically-informed comparisons. The adopted database format and software architecture ensure a fast and efficient analysis and comparison of any pair/set of instruments, in a real-time fashion. In this manuscript, the Python-based PIANO-lab software package is presented. Thanks to a set of acoustic descriptors, PIANO-lab provides a compact and intuitive representation of the Fazioli grand pianos in terms of their spectral, time decay and sound level properties. Based on the implemented data models and graphical tools, even more informative metrics are made available to assess the quality of the instruments produced.