Detecting overblown flute fingerings from the residual noise spectrum

Abstract

Producing a tone by increasing the blowing pressure to excite a higher frequency impedance minimum, or overblowing, is widely used in standard flute technique. In this paper, the effect of overblowing a fingering is explored with spectral analysis, and a fingering detector is designed based on acoustical knowledge and pattern classification techniques. The detector performs signal analysis of the strong broadband signal, that is, spectrally shaped by the pipe impedance, and measures the spectral energy during the attack around multiples of the fundamental frequency sub-multiples over the first octave and a half. It is trained and evaluated on sounds recorded with four expert performers. They played six series of tones from overblown and regular fingerings, with frequencies that are octave- and non-octave-related to the playing frequency. The best of the four proposed sound descriptors allows for a detection error below 1.3% for notes with two and three fingerings (C(5), D(5), C(6), and Cmusical sharp(6)) and below 14% for four (E(6)) or five fingerings (G(6)). The error is shown to dramatically increase when two fingerings' impedance become too similar (E(6) and A(4) and G(6) and C(5)).

Detecting fingering of overblown flute sound using sparse feature learning

“…We created a dataset that resembles the one in a previous study by Verfaille et al [10]. This includes diverse overblown fingering cases such as octave-related to non-octave-related fingerings and "few keys changing" to "many keys changing" positions, as illustrated in Table 1.…”

“…Although we used the same fingering set as [10], our data is recorded without attaching a microphone to the flute head joint. To simulate a real-world situation, audio samples were recorded using the built-in microphone of a Samsung NT900 laptop rather than by using high-end studio microphones.…”

“…This approach uses energy measurements of multiples of the fundamental frequency submultiples F0/l where (l = {2,3,4,5,6}) in the first octave and a half (i.e., between F0 and 1.5F0). This is where, in their observation, the most noticeable differences appear [10]. The spectrum energy is measured using a Hann window centered on the region of interest.…”

“…A study by Kereliuk et al [9] and Verfaille et al [10] attempted to detect overblown flute fingering using the residual noise spectrum with principal component analysis (PCA) and linear discriminant analysis (LDA). This approach uses energy measurements of multiples of the fundamental frequency submultiples F0/l where (l = {2,3,4,5,6}) in the first octave and a half (i.e., between F0 and 1.5F0).…”

Statistical models for the indirect acquisition of violin bowing controls from audio analysis

“…8 Perceptual models cover methods that exploit the knowledge about the influence of instrumental controls on the characteristics of the sound produced. For instance, overblown flute fingerings have been estimated from analysis of the residual noise spectrum 9 and violin controls such as played string, type of excitation (bowed or pizzicato) and the location of the excitation have been extracted from spectral analysis. 10 Finally, statistical learning is composed by methods based on data mining of previously recorded performance data, as it is done by Kereliuk 11 where authors are able to extract flute fingerings corresponding to harmonic notes.…”