Acoustic Analysis of Musical Timbre of Wooden Aerophones

The characterization of the musical timbre, which allows the quantitative evaluation of audios, is still an open-ended research topic. This paper evaluates a set of dimensionless descriptors for studying musical timbre in monophonic recordings of woodwind instruments from the TinySOl audio library, considering the region of frequencies common to all instruments in their three dynamic levels (pianissimo, mezzo-forte, and fortissimo). These descriptors are calculated using the spectra obtained from the Fast Fourier Transform (FFT) using the Python programming language. From the analysis of the distribution of the coefficients, it was possible to verify that the Affinity coefficient (A) allows discrimination in all octaves of musical sounds. The analysis of the data through the empirical distribution of the coefficients shows that the timbral variations due to the dynamics are reflected through the coefficients Sharpness (S) and Mean Affinity (MA). The coefficients are examined using the Principal Component Analysis (PCA), and it was observed that the variability in the distribution of the first principal component is mainly due to the Sharpness (S) and Mean Contrast (MC) coefficients (similar to 55%), and in similar to 43% by the Affinity coefficient. Similarly, the variation in the second principal component is due to 62 % of the MC coefficient and 49% due to MA. It is concluded that the proposed descriptors are sufficient to differentiate the aerophones studied by octaves and musical dynamics.