Recurrence Quantification Analysis of Crowd Sound Dynamics

When multiple individuals interact in a conversation or as part of a large crowd, emergent structures and dynamics arise that are behavioral properties of the interacting group rather than of any individual member of that group. Recent work using traditional signal processing techniques and machine learning has demonstrated that global acoustic data recorded from a crowd at a basketball game can be used to classify emergent crowd behavior in terms of the crowd's purported emotional state. We propose that the description of crowd behavior from such global acoustic data could benefit from nonlinear analysis methods derived from dynamical systems theory. Such methods have been used in recent research applying nonlinear methods to audio data extracted from music and group musical interactions. In this work, we used nonlinear analyses to extract features that are relevant to the behavioral interactions that underlie acoustic signals produced by a crowd attending a sporting event. We propose that recurrence dynamics measured from these audio signals via recurrence quantification analysis (RQA) reflect information about the behavioral dynamics of the crowd itself. We analyze these dynamics from acoustic signals recorded from crowds attending basketball games, and that were manually labeled according to the crowds' emotional state across six categories: angry noise, applause, cheer, distraction noise, positive chant, and negative chant. We show that RQA measures are useful to differentiate the emergent acoustic behavioral dynamics between these categories, and can provide insight into the recurrence patterns that underlie crowd interactions.

The "Morning Voice": The Effect of 24 Hours of Sleep Deprivation on Vocal Parameters of Young Adults

Fatigue following sleep deprivation adversely affects various aspects of human performance. It also induces recognizable voice changes, but the literature is inconsistent regarding their nature. The current study used acoustical analyses to assess the effect of 24 hours of sleep deprivation on vocal parameters of young adults. Forty-seven participants (23 females and 24 males) were tested after nocturnal sleep and after 24 hours of sleep deprivation. Different voice samples were recorded (sustained phonation, words, and sentences) and analyzed for fundamental frequency (F₀, in Hz), vocal intensity (in dB), harmonic-noise ratio (HNR, in dB), jitter, and shimmer (in %). The main finding was significantly higher HNR values following sleep deprivation than nocturnal sleep for females, across all voice samples. The HNR is a measure of the degree of acoustic periodicity, or the amount of noise compared to the harmonic quality present in the voice. As its values are higher, the voice quality is perceptually better. The current results indicate that females had a significantly higher ratio of vocal harmonics compared to vocal noise when they were sleep deprived. In contrast, following nocturnal sleep, the ratio of vocal harmonics compared to vocal noise was lower, which means the voice quality was poorer. This may explain the common perceptual impressions of decreased voice quality after sleep, which may be more pronounced in females.

[Pulmonary crackles, what does the clinician hear?]

INTRODUCTION

The overall duration of a pulmonary crackle is usually less than 20-30 ms but psychoacoustics demonstrates that an acoustical event with a duration of less than 20-40 ms cannot be estimated in terms of pitch and duration. We pose the hypothesis that the main resonant information is contained into the breath sounds following the crackle.

METHODS

Eight patients with COPD, viral pneumonia, bronchiectasis, congestive heart failure, hypoproteinemia and fibrosing alveolitis were recruited for this study. Thirty-six crackles were analyzed in time and frequency domains; 12 in each category of low, medium and high frequencies. The acoustic features of the crackles, their segments (initial deflection width, first cycle duration, two cycles duration, decay segment) and the breath sounds following the crackles were compared.

RESULTS

The study confirms the differences between the three crackles categories in time and frequency domains. No statistical differences were found between the decay segments and breath sounds in each category.

CONCLUSIONS

Breath sounds modified by lung tissue density could be the main resonators determining the fundamental transmission frequencies of crackle signals. Combined acoustic analysis of crackles and breath sounds could replace single analysis of isolated crackles.

Frequency measurement of vowel formants produced by Brazilian children aged between 4 and 8 years

OBJECTIVE

To investigate frequency measurement of the first three formants of the seven oral Brazilian Portuguese vowels of healthy children aged 4-8 years.

METHODS

Two hundred seven children of both genders were selected by oral expression screening and perceptive-auditory analysis. They were separated into four age groups (G1, G2, G3, and G4) and by gender. The voice signals were obtained from key sentence utterance and segments of the seven Brazilian Portuguese oral vowels in tonic position used to estimate formant frequency measurement. Software Praat was used to for processing the recordings.

RESULTS

Findings were presented by mean values of each of the investigated parameters. A rate of 61.90% of statistically significant differences between genders was found, and when analyzing the age groups and genders, we observed that 65 of the 84 items studied (seven vowels × three formant frequencies × four groups) had higher frequencies of formants for girls. There was a decrease in the frequencies values of the first three formants with age. The results recommended grouping of G1 and G2, and they showed a clear difference between this new formed group and G4. In the age groups of 5-year old to 6 years 11 months (G2 and G3) and 6-year old to 7 years 11 months (G3 and G4), there were statistically significant changes that were random for parameter and vowel. There was a decrease in the frequencies values of the first three formants with age.

CONCLUSION

Formant frequencies showed a tendency to differentiate genders and their absolute values were in general higher in girls. Age increases showed decreases in formant frequencies. Tests for statistical differences led to grouping of G1 and G2 and a clear difference between this new formed group and G4. The comparison between G2 and G3 and G3 and G4 showed random changes. The changes during this age period (5-year old to 7 years 11 months) were attributed to a transition stage of acoustic measurements in children. As formant frequencies vary according to structural and postural aspects of the vocal tract and speech organs, their study in healthy children contributes for the understanding of the development of the pediatric phonation system, in addition to offering a reference data set for future studies of children with vocal disorders that can potentially impact the resonance system.

Perceptual and acoustic analysis of voice in individuals with total thyroidectomy: pre-post surgery comparison [corrected]

Total thyroidectomy (TT) is a surgical procedure which involves complete removal of the thyroid gland, usually done in massive goitre compressing the trachea and esophagus, hyperthyroidism and carcinoma of thyroid gland. Laryngeal nerve damage is one of the most feared complications after TT which can lead to permanent changes in voice. Recent research suggests the occurrence of voice changes without any visible laryngeal nerve damage. Present study attempts to compare the pre and post-operative voice characteristics in individuals with total thyroidectomy without any laryngeal nerve damage. A total of 27 subjects (21 females and 6 males) who underwent total thyroidectomy participated in the study. Their recorded phonation of vowel/a/was subjected to two types of analyses viz. Perceptual analysis (using CAPE-V rating scale) and Acoustic analysis (using Multi Dimensional Voice Profile). Results of perceptual analysis indicated slight decrease in overall severity, roughness and breathiness and a slight increase in strainness, in only males. Acoustic analysis findings supported perceptual results with minimal changes in most of the parameters. The results suggest that after total thyroidectomy, in the absence of visible laryngeal nerve damage, functional changes in voice are minimal and temporary in nature. This study provides an insight to Otolaryngologists and Speech Language Pathologists about the voice characteristics in individuals with thyroidectomy, enabling them to formulate appropriate therapy protocol for this population. It further attempts to sensitize surgeons and physicians on the need for referral of this population to Speech Language Pathologist in the event of dysphonia.