We present a framework for the acquisition and parametrization of object material properties. The introduced acquisition device, denoted as Texplorer2, is able to extract surface material properties while a human operator is performing exploratory procedures. Using the Texplorer2, we scanned 184 material classes which we labeled according to biological, chemical, and geological naming conventions. Based on these real material recordings, we introduce a novel set of mathematical features which align with corresponding material properties defined in perceptual studies from related work and classify the materials using common machine learning techniques. Validation results of the proposed multi-modal features lead to an overall classification accuracy of 90.2% +/- 1.2% and an F1 score of 0.90 +/- 0.01 using a random forest classifier. For the sake of comparison, a deep neural network is trained and tested on images of the surfaces; it outperforms (91.3% +/- 0.7%) the hand-crafted feature-based approach yet leads to more critical misclassifications in terms of the proposed taxonomy.

Within the realm of voice classification, singers could be sub-categorized by the weight of their repertoire, the so-called “singer's Fach.” However, the opposite pole terms “lyric” and “dramatic” singing are not yet well defined by their acoustic and articulatory characteristics. Nine professional singers of different singers' Fach were asked to sing a diatonic scale on the vowel /a/, first in what the singers considered as lyric and second in what they considered as dramatic. Image recording was performed using real time magnetic resonance imaging (MRI) with 25 frames/s, and the audio signal was recorded via an optical microphone system. Analysis was performed with regard to sound pressure level (SPL), vibrato amplitude, and frequency and resonance frequencies as well as articulatory settings of the vocal tract. The analysis revealed three primary differences between dramatic and lyric singing: Dramatic singing was associated with greater SPL and greater vibrato amplitude and frequency as well as lower resonance frequencies. The higher SPL is an indication of voice source changes, and the lower resonance frequencies are probably caused by the lower larynx position. However, all these strategies showed a considerable individual variability. The singers' Fach might contribute to perceptual differences even for the same singer with regard to the respective repertoire.

The human voice exhibits diverse expressions, primarily generated by the regular oscillatory movement of both vocal folds. However, further vocal tract structures can also oscillate, as observed in different singing styles, supraglottic voice substitutes and functional voice disorders. To describe the spectrum of irregular phonation types from a nonlinear systems perspective, we conducted in vivo investigations with a professional metal singer using transnasal high-speed digital imaging combined with electroglottography and acoustic data. Here, we analysed 10 distinct voice qualities. Some of these show oscillatory mechanisms characterized by nearly periodic patterns that differ in three key aspects: the ratio of glottal to supraglottic oscillations, their behaviour during the glissando and the oscillating structures primarily involved. Both reduced glottal adduction and increased glottal and supraglottic compression contribute to a greater degree of irregularity. The supraglottic oscillations can coexist with various glottal oscillation mechanisms. Based on the in vivo data, we identified various nonlinear phenomena of different categories that include synchronized oscillations with defined frequency ratios, bifurcations, biphonations and chaotic oscillations.

Introduction: Despite its importance in voice training, comprehensive research into sustained vowel phonation with constant pitch and increasing and decreasing loudness, the so-called Messa di Voce, is lacking. The study examines the laryngeal behavior during Messa di Voce, regarding the impact of the speed of execution on voice stability parameters. Materials and methods: Nine untrained, healthy subjects (5 female, 4 male) were asked to perform Messa di Voce exercises on the vowel [i:], involving a gradual increase and decrease of volume. During the first task, each phase should take 3 s, whereas in the second task, each phase should take 1 s. Female subjects sang pitch B3 (fundamental frequency f_o ≈ 247 Hz), and male subjects pitch B2 (f_o ≈ 124 Hz). Throughout phonation, synchronous recordings were captured through high-speed videolaryngoscopy (HSV), electroglottography, and audio signals. Subsequently, the Glottal Area Waveform was extracted from the HSV data. The tasks’ duration and calculated parameters (including, e.g., Open Quotient (OQ), Closing Quotient (ClQ), Relative Average Perturbation (RAP)), excluding parts of the signal with stationary sound pressure level (SPL), were analyzed with correlation analysis and statistical analysis (Analysis of Variance and subsequent multiple comparisons). Results: Subjects shortened the requested task length by factor ≈ 0.5. The f_o remained almost stable for most subjects and tasks. There were strong negative correlations between SPL and both OQ and ClQ. The median RAP appears to decrease towards the SPL apex and then increase again. Statistical effects were shown especially for females during the fast task, which may be due to raised SPL. Conclusion: There was no specific effect on stability found corresponding to the task’s speed. Also, no major vocal instabilities at a specific sound pressure level were apparent, indicating no transitions as they exist for f_o regions with registration events.

This study aimed to characterize the biomechanics of vibrato and non-vibrato phonation in professional singers, focusing on vocal fold oscillation patterns and cuneiform cartilage movements. Ten vocally healthy, professionally-trained, western classical singers were instructed to sustain the vowel [i:] without vibrato, with vibrato, and again without vibrato, for approximately two seconds for each quality. Data were collected via acoustic recordings, electroglottography (EGG), and transnasal high-speed videoendoscopy. After segmentation, glottal area waveforms (GAW) were analyzed, and temporal trajectories of fundamental frequency ( fo) and glottal parameters were compared. Cuneiform cartilage movement was tracked in three spatial dimensions using the videoendoscopic image sequences. The participants maintained a relatively stable mean fo across conditions, with no significant differences in open quotient (OQEGG and OQGAW) and closing quotient (ClQ) regarding vibrato and non-vibrato. During vibrato, most participants showed a higher OQ in vibrato fo minima than in maxima. Vibrato was associated with lower cepstral peak prominence and higher sound pressure level (SPL), and SPL courses inverse to fo modulation. The cuneiform tubercle tracking correlated to fo showed vertical and horizontal movements which followed, in one subject, an epiglottis forward positioning, presumably due to a tilting by cricothyroid muscle activity linked to the vibrato phenomenon.