Neural network-based estimation of biomechanical vocal fold parameters.

Vocal fold (VF) vibrations are the primary source of human phonation. High-speed video (HSV) endoscopy enables the computation of descriptive VF parameters for assessment of physiological properties of laryngeal dynamics, i.e., the vibration of the VFs. However, underlying biomechanical factors responsible for physiological and disordered VF vibrations cannot be accessed. In contrast, physically based numerical VF models reveal insights into the organ's oscillations, which remain inaccessible through endoscopy. To estimate biomechanical properties, previous research has fitted subglottal pressure-driven mass-spring-damper systems, as inverse problem to the HSV-recorded VF trajectories, by global optimization of the numerical model. A neural network trained on the numerical model may be used as a substitute for computationally expensive optimization, yielding a fast evaluating surrogate of the biomechanical inverse problem. This paper proposes a convolutional recurrent neural network (CRNN)-based architecture trained on regression of a physiological-based biomechanical six-mass model (6 MM). To compare with previous research, the underlying biomechanical factor "subglottal pressure" prediction was tested against 288 HSV ex vivo porcine recordings. The contributions of this work are two-fold: first, the presented CRNN with the 6 MM handles multiple trajectories along the VFs, which allows for investigations on local changes in VF characteristics. Second, the network was trained to reproduce further important biomechanical model parameters like VF mass and stiffness on synthetic data. Unlike in a previous work, the network in this study is therefore an entire surrogate of the inverse problem, which allowed for explicit computation of the fitted model using our approach. The presented approach achieves a best-case mean absolute error (MAE) of 133 Pa (13.9%) in subglottal pressure prediction with 76.6% correlation on experimental data and a re-estimated fundamental frequency MAE of 15.9 Hz (9.9%). In-detail training analysis revealed subglottal pressure as the most learnable parameter. With the physiological-based model design and advances in fast parameter prediction, this work is a next step in biomechanical VF model fitting and the estimation of laryngeal kinematics.

Analyzing Vocal Fold Frequency Dynamics Using High-Speed 3D Laser Video Endoscopy.

OBJECTIVE: To examine changes in lateral and vertical vibratory motion along the anterior, middle, and posterior sections of the vocal folds, as a function of vocal frequency variations. METHODS: Absolute measurements of vocal fold surface dynamics from high-speed videoendoscopy with custom laser endoscope were made on 23 vocally healthy adults during sustained /i:/ production at 10%, 20%, and 80% of pitch range. The 3D parameters of amplitude (mm), maximum velocity opening/closing (mm/s), and mean velocity opening/closing (mm/s) were computed for the lateral and vertical vibratory motion along the anterior, middle, and posterior sections of the vocal folds. Linear mixed model analysis was conducted to evaluate the differences in (a) vocal frequency levels (high vs. normal vs. low pitch), (b) axis level (vertical vs. lateral), (c) position level (anterior vs. middle vs. posterior), and (d) gender differences (male vs. female). RESULTS: Overall, the superior surface vertical motion of the vocal fold is greater compared with the lateral motion, especially in males. Along the superior surface, the mean and maximum closing velocities are greater posteriorly for low pitch. The location (anterior, middle, and posterior) along the superior surface is relevant only for vocal fold closing rather than opening, as the dynamics are different along the various locations. CONCLUSIONS: The study highlights the significance of assessing the vertical motion of the superior surface of the vocal fold to understand the complex dynamics of voice production. LEVEL OF EVIDENCE: NA Laryngoscope, 2024.

Machine learning based estimation of hoarseness severity using sustained vowelsa).

Auditory perceptual evaluation is considered the gold standard for assessing voice quality, but its reliability is limited due to inter-rater variability and coarse rating scales. This study investigates a continuous, objective approach to evaluate hoarseness severity combining machine learning (ML) and sustained phonation. For this purpose, 635 acoustic recordings of the sustained vowel /a/ and subjective ratings based on the roughness, breathiness, and hoarseness scale were collected from 595 subjects. A total of 50 temporal, spectral, and cepstral features were extracted from each recording and used to identify suitable ML algorithms. Using variance and correlation analysis followed by backward elimination, a subset of relevant features was selected. Recordings were classified into two levels of hoarseness, H<2 and H≥2, yielding a continuous probability score y∈[0,1]. An accuracy of 0.867 and a correlation of 0.805 between the model's predictions and subjective ratings was obtained using only five acoustic features and logistic regression (LR). Further examination of recordings pre- and post-treatment revealed high qualitative agreement with the change in subjectively determined hoarseness levels. Quantitatively, a moderate correlation of 0.567 was obtained. This quantitative approach to hoarseness severity estimation shows promising results and potential for improving the assessment of voice quality.

Machine Learning-assisted immunophenotyping of peripheral blood identifies innate immune cells as best predictor of response to induction chemo-immunotherapy in head and neck squamous cell carcinoma - knowledge obtained from the CheckRad-CD8 trial.

PURPOSE: Individual prediction of treatment response is crucial for personalized treatment in multimodal approaches against head-and-neck squamous cell carcinoma (HNSCC). So far, no reliable predictive parameters for treatment schemes containing immunotherapy have been identified. This study aims to predict treatment response to induction chemo-immunotherapy based on the peripheral blood immune status in patients with locally advanced HNSCC. METHODS: The peripheral blood immune phenotype was assessed in whole blood samples in patients treated in the phase II CheckRad-CD8 trial as part of the pre-planned translational research program. Blood samples were analyzed by multicolor flow cytometry before (T1) and after (T2) induction chemo-immunotherapy with cisplatin/docetaxel/durvalumab/tremelimumab. Machine Learning techniques were used to predict pathological complete response (pCR) after induction therapy. RESULTS: The tested classifier methods (LDA, SVM, LR, RF, DT, and XGBoost) allowed a distinct prediction of pCR. Highest accuracy was achieved with a low number of features represented as principal components. Immune parameters obtained from the absolute difference (lT2-T1l) allowed the best prediction of pCR. In general, less than 30 parameters and at most 10 principal components were needed for highly accurate predictions. Across several datasets, cells of the innate immune system such as polymorphonuclear cells, monocytes, and plasmacytoid dendritic cells are most prominent. CONCLUSIONS: Our analyses imply that alterations of the innate immune cell distribution in the peripheral blood following induction chemo-immuno-therapy is highly predictive for pCR in HNSCC.

Reconstruction of Vocal Fold Medial Surface 3D Trajectories: Effects of Neuromuscular Stimulation and Airflow.

INTRODUCTION: Analysis of medial surface dynamics of the vocal folds (VF) is critical to understanding voice production and treatment of voice disorders. We analyzed VF medial surface vibratory dynamics, evaluating the effects of airflow and nerve stimulation using 3D reconstruction and empirical eigenfunctions (EEF). STUDY DESIGN: In vivo canine hemilarynx phonation. METHODS: An in vivo canine hemilarynx was phonated while graded stimulation of the recurrent and superior laryngeal nerves (RLN and SLN) was performed. For each phonatory condition, vibratory cycles were 3D reconstructed from tattooed landmarks on the VF medial surface at low, medium, and high airflows. Parameters describing medial surface trajectory shape were calculated, and underlying patterns were emphasized using EEFs. Fundamental frequency and smoothed cepstral peak prominence (CPPS) were calculated from acoustic data. RESULTS: Convex-hull area of landmark trajectories increased with increasing flow and decreasing nerve activation level. Trajectory shapes observed included circular, ellipsoid, bent, and figure-eight. They were more circular on the superior and anterior VF, and more elliptical and line-like on the inferior and posterior VF. The EEFs capturing synchronal opening and closing (EEF1) and alternating convergent/divergent (EEF2) glottis shapes were mostly unaffected by flow and nerve stimulation levels. CPPS increased with higher airflow except for low RLN activation and very dominant SLN stimulation. CONCLUSION: We analyzed VF vibration as a function of neuromuscular stimulation and airflow levels. Oscillation patterns such as figure-eight and bent trajectories were linked to high nerve activation and flow. Further studies investigating longer sections of 3D reconstructed oscillations are needed. LEVEL OF EVIDENCE: N/A, Basic Science Laryngoscope, 134:1249-1257, 2024.

Bilateral Functional Electrical Stimulation for the Treatment of Presbyphonia in a Sheep Model.

OBJECTIVES: The aim of the study was to increase muscle volume and improve phonation characteristics of the aged ovine larynx by functional electrical stimulation (FES) using a minimally invasive surgical procedure. METHODS: Stimulation electrodes were placed bilaterally near the terminal adduction branch of the recurrent laryngeal nerves (RLN). The electrodes were connected to battery powered pulse generators implanted subcutaneously at the neck region. Training patterns were programmed by an external programmer using a bidirectional radio frequency link. Training sessions were repeated automatically by the implant every other day for 1 week followed by every day for 8 weeks in the awake animal. Another group of animals were used as sham, with electrodes positioned but not connected to an implant. Outcome parameters included gene expression analysis, histological assessment of muscle fiber size, functional analysis, and volumetric measurements based on three-dimensional reconstructions of the entire thyroarytenoid muscle (TAM). RESULTS: Increase in minimal muscle fiber diameter and an improvement in vocal efficiency were observed following FES, compared with sham animals. CONCLUSION: This is the first study to demonstrate beneficial effects in the TAM of FES at molecular, histological, and functional levels. FES of the terminal branches of the RLN reversed the effects of age-related changes and improved vocal efficiency. LEVEL OF EVIDENCE: NA Laryngoscope, 134:848-854, 2024.

Computational fluid dynamics of upper airway aerodynamics for exercise-induced laryngeal obstruction: A feasibility study.

Objective: Use of computational fluid dynamic (CFD) simulations to measure the changes in upper airway geometry and aerodynamics during (a) an episode of Exercise-Induced Laryngeal Obstruction (EILO) and (b) speech therapy exercises commonly employed for patients with EILO. Methods: Magnetic resonance imaging stills of the upper airway including the nasal and oral cavities from an adult female were used to re-construct three-dimensional geometries of the upper airway. The CFD simulations were used to compute the maximum volume flow rate (l/s), pressure (Pa), airflow velocity (m/s) and area of cross-section opening in eight planes along the vocal tract, separately for inhalation and exhalation. Results: Numerical predictions from three-dimensional geometrical modeling of the upper airway suggest that the technique of nose breathing for inhalation and pursed lip breathing for exhalation show most promising pressure conditions and cross-sectional diameters for rescue breathing exercises. Also, if EILO is due to the constriction at the vocal fold level, then a quick sniff may also be a proper rescue inhalation exercise. EILO affects both the inspiratory and the expiratory phases of breathing. Conclusions: A prior knowledge of the supraglottal aerodynamics and the corresponding upper airway geometry from CFD analysis has the potential to assist the clinician in choosing the most effective rescue breathing technique for optimal functional outcome of speech therapy intervention in patients with EILO and in understanding the pathophysiology of EILO on a case-by-case basis with future studies. Level of Evidence: 4.

Validation and enhancement of a vocal fold medial surface 3D reconstruction approach for in-vivo application.

In laryngeal research, studying the vertical vocal fold oscillation component is often disregarded. However, vocal fold oscillation by its nature is a three-dimensional process. In the past, we have developed an in-vivo experimental protocol to reconstruct the full, three-dimensional vocal fold vibration. The goal of this study is to validate this 3D reconstruction method. We present an in-vivo canine hemilarynx setup using high-speed video recording and a right-angle prism for 3D reconstruction of vocal fold medial surface vibrations. The 3D surface is reconstructed from the split image provided by the prism. For validation, reconstruction error was calculated for objects located at a distance of up to 15 mm away from the prism. The influence of camera angle, changing calibrated volume, and calibration errors were determined. Overall average 3D reconstruction error is low and does not exceed 0.12 mm at 5 mm distance from the prism. Influence of a moderate (5°) and large (10°) deviation in camera angle led to a slight increase in error to 0.16 mm and 0.17 mm, respectively. This procedure is robust towards changes in calibration volume and small calibration errors. This makes this 3D reconstruction approach a useful tool for the reconstruction of accessible and moving tissue surfaces.

Effect of functional electric stimulation on phonation in an ex vivo aged ovine model.

With age, the atrophy of the thyroarytenoid muscle (TAM), and thus atrophy of the vocal folds, leads to decreased glottal closure, increased breathiness, and a loss in voice quality, which results in a reduced quality of life. A method to counteract the atrophy of the TAM is to induce hypertrophy in the muscle by functional electric stimulation (FES). In this study, phonation experiments were performed with ex vivo larynges of six stimulated and six unstimulated ten-year-old sheep to investigate the impact of FES on phonation. Electrodes were implanted bilaterally near the cricothyroid joint. FES treatment was provided for nine weeks before harvesting. The multimodal measurement setup simultaneously recorded high-speed video of the vocal fold oscillation, the supraglottal acoustic signal, and the subglottal pressure signal. Results of 683 measurements show a 65.6% lower glottal gap index, a 22.7% higher tissue flexibility (measured by the amplitude to length ratio), and a 473.7% higher coefficient of determination (R2) of the regression of subglottal and supraglottal cepstral peak prominence during phonation for the stimulated group. These results suggest that FES improves the phonatory process for aged larynges or presbyphonia.

Nyquist Plot Parametrization for Quantitative Analysis of Vibration of the Vocal Folds.

OBJECTIVES: The Nyquist plot provides a graphical representation of the glottal cycles as elliptical trajectories in a 2D plane. This study proposes a methodology to parameterize the Nyquist plot with application to support the quantitative analysis of voice disorders. METHODS: We considered high-speed videoendoscopy recordings of 33 functional dysphonia (FD) patients and 33 normophonic controls (NC). Quantitative analysis was performed by computing four shape-based parameters from the Nyquist plot: Variability, Size (Perimeter and Area), and Consistency. Additionally, we performed automatic classification using a linear support vector machine and feature importance analysis by combining the proposed features with state-of-the-art glottal area waveform (GAW) parameters. RESULTS: We found that the inter-cycle variability was significantly higher in FD patients compared to NC. We achieved a classification accuracy of 83% when the top 30 most important features were used. Furthermore, the proposed Nyquist plot features were ranked in the top 12 most important features. CONCLUSIONS: The Nyquist plot provides complementary information for subjective and objective assessment of voice disorders. On the one hand, with visual inspection it is possible to observe intra- and inter-glottal cycle irregularities during sustained phonation. On the other hand, shaped-based parameters allow quantifying such irregularities and provide complementary information to state-of-the-art GAW parameters.

GlottisNetV2: Temporal Glottal Midline Detection Using Deep Convolutional Neural Networks.

High-speed videoendoscopy is a major tool for quantitative laryngology. Glottis segmentation and glottal midline detection are crucial for computing vocal fold-specific, quantitative parameters. However, fully automated solutions show limited clinical applicability. Especially unbiased glottal midline detection remains a challenging problem. We developed a multitask deep neural network for glottis segmentation and glottal midline detection. We used techniques from pose estimation to estimate the anterior and posterior points in endoscopy images. Neural networks were set up in TensorFlow/Keras and trained and evaluated with the BAGLS dataset. We found that a dual decoder deep neural network termed GlottisNetV2 outperforms the previously proposed GlottisNet in terms of MAPE on the test dataset (1.85% to 6.3%) while converging faster. Using various hyperparameter tunings, we allow fast and directed training. Using temporal variant data on an additional data set designed for this task, we can improve the median prediction accuracy from 2.1% to 1.76% when using 12 consecutive frames and additional temporal filtering. We found that temporal glottal midline detection using a dual decoder architecture together with keypoint estimation allows accurate midline prediction. We show that our proposed architecture allows stable and reliable glottal midline predictions ready for clinical use and analysis of symmetry measures.

Influence of Loudness on Vocal Stability in the Male Passaggio.

INTRODUCTION: Vocal registers and the frequency region where registration events occur, the passaggio, have been in focus of scientific research for almost 200 years. In professional tenors, it has been shown before that singing across the passaggio avoiding a register shift and therefore using their stage voice above the passaggio (SVaP) is associated with greater vocal stability than a register change to the falsetto. However, it is unclarified how much different loudness conditions contribute to this vocal stability. MATERIAL AND METHODS: Six professional tenors were asked to perform four pitch glides from A3 to A4 (220-440 Hz) on the vowel [i:]. These glides included (1) the passaggio from modal register to falsetto. The following glides into SVaP were performed under different loudness conditions, (2) mezzoforte (average loudness), (3) pianissimo (as quietly as possible), and (4) fortissimo (the loudest possible). During phonation, high speed videoendoscopy (HSV), electroglottography, and audio signals were recorded simultaneously. The glottal area waveform was derived based on the HSV material. RESULTS: Modal to falsetto transitions were associated with relatively low sound pressure level and rise of open quotients (OQ) for the falsetto. Transitions to SVaP showed a clear dependence on the intended loudness. The OQs were lower the louder the task was. There was no clear evidence that transitions with softer voice showed greater stability of vocal fold oscillation patterns than louder tasks. CONCLUSIONS: The vocal fold oscillation pattern show- differences among various loudness conditions within the tenors' passaggio but no clear differences with regard to oscillatory stability.

Sources of Aerosol Dispersion During Singing and Potential Safety Procedures for Singers.

INTRODUCTION: With respect to the Covid-19 pandemic, singing is assumed to be associated with a high potential person-to-person transmission. However, it remains unclear how the impulse dispersion varies with different types of articulation, intensity levels of diction, or body position. Furthermore, it has not been understood in detail how to prevent aerosol dispersion during singing. MATERIAL AND METHODS: Single professional singers from the Bavarian Radio Chorus were asked to sing in different head positions, with different articulation patterns and different masks after inhaling the basic liquid of an e-cigarette. The vapor cloud was segmented and tracked over time. RESULTS: Consonants and exaggeration of diction enhanced the distance reached by the impulse dispersion. Furthermore, the greatest dispersion was reached for a neutral head position. All protection masks stopped the initial jet of the aerosols but the FFP2 masks were the most effective. CONCLUSION: Some protection equipment has been identified to be promising in reducing aerosol dispersion. However, systematic effects have to be evaluated in greater collectives.

Synthetic mucus for an ex vivo phonation setup: Creation, application, and effect on excised porcine larynges.

Laryngeal mucus hydrates and lubricates the deformable tissue of the vocal folds and acts as a boundary layer with the airflow from the lungs. However, the effects of the mucus' viscoelasticity on phonation remain widely unknown and mucus has not yet been established in experimental procedures of voice research. In this study, four synthetic mucus samples were created on the basis of xanthan with focus on physiological frequency-dependent viscoelastic properties, which cover viscosities and elasticities over 2 orders of magnitude. An established ex vivo experimental setup was expanded by a reproducible and controllable application method of synthetic mucus. The application method and the suitability of the synthetic mucus samples were successfully verified by fluorescence evidence on the vocal folds even after oscillation experiments. Subsequently, the impact of mucus viscoelasticity on the oscillatory dynamics of the vocal folds, the subglottal pressure, and acoustic signal was investigated with 24 porcine larynges (2304 datasets). Despite the large differences of viscoelasticity, the phonatory characteristics remained stable with only minor statistically significant differences. Overall, this study increased the level of realism in the experimental setup for replication of the phonatory process enabling further research on pathological mucus and exploration of therapeutic options.

Error detection and filtering of incompressible flow simulations for aeroacoustic predictions of human voice.

The presented filtering technique is proposed to detect errors and correct outliers inside the acoustic sources, respectively, the first time derivative of the incompressible pressure obtained from large eddy simulations with prescribed vocal fold motion using overlay mesh methods. Regarding the perturbed convective wave equation, the time derivative of the incompressible pressure is the primary sound source in the human phonation process. However, the incompressible pressure can be erroneous and have outliers when fulfilling the divergence-free constraint of the velocity field. This error is primarily occurring for non-conserving prescribed vocal fold motions. Therefore, the method based on a continuous stationary random process was designed to detect rare events in the time derivative of the pressure. The detected events are then localized and treated by a defined window function to increase their probability. As a consequence, the data quality of the non-linearly filtered data is enhanced significantly. Furthermore, the proposed method can also be used to assess convergence of the aeroacoustic source terms, and detect regions and time intervals, which show a non-converging behavior by an impulse-like structure.

Impulse dispersion of aerosols during playing the recorder and evaluation of safety measures.

INTRODUCTION: Group musical activities using wind instruments have been restricted during the CoVID19 pandemic due to suspected higher risk of virus transmission. It was presumed that the aerosols exhaled through the tubes while playing would be ejected over larger distances and spread into the room due to jet stream effects. In particular, the soprano recorder is widely used as an instrument in school classes, for beginners of all age groups in their musical education, in the context of leisure activities and in professional concert performances. Understanding the aerosol impulse dispersion characteristics of playing the soprano recorder could assist with the establishment of concepts for safe music-making. METHODS: Five adult professionally trained soprano recorder players (4 female, 1 male) played four bars of the main theme of L. van Beethoven's "Ode to Joy" in low and in high octaves, as well as with 3 different potential protection devices in the high octave. For comparison they spoke the corresponding text by F. Schiller. Before each task, they inhaled .5 L of vapor from an e-cigarette filled with base liquid. The vapor cloud escaping during speaking or playing was recorded by cameras and its spread was measured as a function of time in the three spatial dimensions. The potential safety devices were rated for practicability with a questionnaire, and their influence on the sound was compared, generating a long-term average spectrum from the audio data. RESULTS: When playing in the high octave, at the end of the task the clouds showed a median distance of 1.06 m to the front and .57 m diameter laterally (maxima: x: 1.35 m and y: .97 m). It was found that the clouds' expansion values in playing the recorder with and without safety measures are mostly lower when compared to the ordinary, raised speaking voice of the same subjects. The safety devices which covered the instrument did not show clear advantages and were rated as unpractical by the subjects. The most effective reduction of the cloud was reached when playing into a suction funnel. CONCLUSION: The aerosol dispersion characteristics of soprano recorders seem comparable to clarinets. The tested safety devices which covered holes of the instrument did not show clear benefits.

The Effect of Water Resistance Therapy on the Impulse Dispersion of Aerosols During Sustained Phonation.

INTRODUCTION: Group singing has been associated with higher transmission risks via exhaled and spread aerosols in the CoVID19 pandemic. For this reason, many musical activities, such as rehearsals and lessons, but also voice therapy sessions, have been restricted in many countries. Consequently, transmission risks and pathways have been studied, such as aerosol amounts generated by exhalation tasks, convectional flows in rooms, or the impulse dispersion of different kinds of phonation. The use of water resistance exercises such as those utilizing LAX VOX®, are common in voice lessons and as vocal warm-ups. With this context, this study investigates the impulse dispersion characteristics of aerosols during a voiced water resistance exercise in comparison to normal singing. METHODS: Twelve professional singers (six male, six female) were asked to phonate a stable pitch through a silicone tube into a bottle filled with water, holding the end of the tube 5 cm below the surface. Before performing the tasks, the singers inhaled the vapor consisting of 0.5 L base liquid from an e-cigarette. The exhaled gas cloud coming out of the bottle was recorded in all three spatial directions and the dispersion was measured as a function of time. RESULTS: At the end of the phonation task, the median distance to the front was 0.55 m and the median of the lateral expansion of the cloud was 0.89 m, the maximum to the front reached 0.88 m, and the maximum of lateral expansion 1.05 m. For the upwards direction of the clouds a median of 1.00 m and a maximum of 1.34 m from the mouth were measured. Three seconds after the end of the task, the medians were declining. CONCLUSION: The exhaled aerosol cloud can expand despite the obstacle of the water when using LAX VOX® during phonation.

Evaluation of Respiratory Particle Emission during Otorhinolaryngological Procedures in the Context of the SARS-CoV-2 Pandemic.

Understanding the risk of infection by routine medical examination is important for the protection of the medical personnel. In this study we investigated respiratory particles emitted by patients during routine otolaryngologic procedures and assessed the risks for the performing physician. We developed two experimental setups to measure aerosol and droplet emission during rigid/flexible laryngoscopy, rhinoscopy, pharyngoscopy, otoscopy, sonography and patient interview for subjects with and without masks. A high-speed-camera setup was used to detect ballistic droplets (approx. > 100 µm) and an aerosol-particle-sizer was used to detect aerosol particles in the range of 0.3 µm to 10 µm. Aerosol particle counts were highly increased for coughing and slightly increased for heavy breathing in subjects without masks. The highest aerosol particle counts occurred during rigid laryngoscopy. During laryngoscopy and rhinoscopy, the examiner was exposed to increased particle emission due to close proximity to the patient's face and provoked events such as coughing. However, even during sonography or otoscopy without a mask, aerosol particles were expelled close to the examiner. The physician's exposure to respiratory particles can be reduced by deliberate choice of examination technique depending on medical indication and the use of appropriate equipment for the examiners and the patients (e.g., FFP2 masks for both).

Impulse dispersion of aerosols during playing wind instruments.

Musical activities, especially singing and playing wind instruments, have been singled out as potentially high-risk activities for the transmission of SARS CoV-2, due to a higher rate of aerosol production and emission. Playing wind instruments can produce condensation, droplets of saliva, and aerosol particles, which hover and spread in the environmental air's convectional flows and which can be potentially infectious. The aim of this study is to investigate the primary impulse dispersion of aerosols that takes place during the playing of different wind instruments as compared to breathing and to speaking. Nine professional musicians (3 trumpeters, 3 flautists and 3 clarinetists) from the Bavarian Symphony Orchestra performed the main theme from the 4th movement of Ludwig van Beethoven's 9th symphony in different pitches and loudness. The inhaled air volume was marked with small aerosol particles produced using a commercial e-cigarette. The expelled aerosol cloud was recorded by cameras from different perspectives. Afterwards, the dimensions and dynamics of the aerosol cloud were measured by segmenting the video footage at every time point. Overall, the flutes produced the largest dispersion at the end of the task, reaching maximum forward distances of 1.88 m. An expulsion of aerosol was observed in different directions: upwards and downwards at the mouthpiece, at the end of the instrument, and along the flute at the key plane. In comparison, the maximum impulse dispersions generated by the trumpets and clarinets were lower in frontal and lateral direction (1.2 m and 1.0 m towards the front, respectively). Also, the expulsion to the sides was lower.