Shenandoah Conservatory’s voice pedagogy graduate program has achieved a record number of accepted papers for presentation at the Voice Foundation’s 51st anniversary symposium “Care of the Professional Voice” in Philadelphia, Pennsylvania, from Wednesday, June 1, through Sunday, June 5. This is the field’s premier international conference and no other voice pedagogy graduate program has as many presentations as Shenandoah.
Seven research projects were accepted for presentation:
- Singing Voice Symptomatology Following Presumed SARS-CoV-2 Infection (Shenandoah University student and faculty collaboration)
- The Relationship Between Speech-Language Pathologists (SLPs) and Young Singers: A Survey (Shenandoah University student and faculty collaboration)
- Vocal Tracts in Virtual Reality – A Preliminary Report (Shenandoah University, University of Iowa and University of London faculty collaboration)
- Physical Modelling of High-Resolution Hybrid CT+MRI Vocal Tract Images (Shenandoah University, University of Iowa and University of London faculty collaboration)
- Minimal CT Dosages Necessary for Blended CT+MRI Vocal Tract Imaging (Shenandoah University, University of Iowa and University of London faculty collaboration)
- Formant/Harmonic Coupling Patterns in Contemporary Belt as Observed in “And I Am Telling You” from Dreamgirls as sung by Amber Riley (Shenandoah University student and faculty collaboration)
- Older African American Singers and Singing Cessation (Shenandoah University student and faculty collaboration)
Shenandoah University Student & Faculty Collaboration
Singing Voice Symptomatology Following Presumed SARS-CoV-2 Infection
Introduction: COVID-19 has changed the world of singing. It is estimated that as many as one in seven persons in the USA have contracted SARS-CoV-2 (coronavirus.jhu.edu, 2021), and as many as 14% of survivors experience ongoing symptoms or “long-COVID” (Page et al., 2021). How these ongoing sequelae impact singers’ vocal function is not well understood. An international survey of singers who have contracted and survived SARS-CoV-2 infection may help the singing voice community better understand the ongoing symptomatology of these singers.
Methods: An online questionnaire will be designed and administered to singers in the Americas, Europe, Australia, and Asia. Survey questions will include demographic information, peri- and post- SARS-CoV-2 infection symptoms, and any self-reported sequelae attributed to long-COVID. The survey will be made available in English, Portuguese, Spanish, and Mandarin Chinese. Data will be analyzed with descriptive statistics to provide a useful summary and description of the sample, and the investigation will be conducted in full compliance with the Shenandoah University Institutional Review Board.
Results/Conclusion: These data will add to our understanding of this growing population’s unique vocal needs, and inform strategies for singing voice habilitation in COVID-19 survivors.
Shenandoah University Student & Faculty Collaboration
The Relationship Between Speech-Language Pathologists (SLPs) and Young Singers: A Survey
Objective: Successful voice care teams are interdisciplinary, including laryngologists, speech language pathologists (SLPs), and singing voice teachers. Each of these persons promote optimal voice function in their clients (Sataloff, 2017), and there may be overlap in their scope of practice. SLPs and singing voice teachers often collaborate on issues related to singers’ vocal health, but students and young singing teachers may be unaware of these beneficial partnerships. This survey will assess the knowledge of Shenandoah Conservatory voice students regarding speech-language pathology.
Methods: An online survey will be generated and distributed electronically to Shenandoah Conservatory students who are enrolled in voice instruction. Questions will investigate students’ awareness of speech language pathologists and their work. Data will be analyzed with descriptive statistics to provide a useful summary and description of the sample, and the investigation will be conducted in full compliance with the Shenandoah University Institutional Review Board.
Results and Conclusion: Data collected will reveal what knowledge Shenandoah Conservatory singers have regarding SLP practice, and if singers feel potential collaborations may be beneficial. If young voice instructors understand the benefits of an interdisciplinary voice care team, they may better serve their future singing voice students. This information will also inform curriculum design, and bridge the knowledge gap between SLPs and singing voice teachers.
Shenandoah University, University of Iowa & University of London Faculty Collaboration
Vocal Tracts in Virtual Reality – A Preliminary Report
Introduction: Virtual reality (VR) and immersive learning environments (ILEs) are used increasingly for efficient training and pedagogy. Providing a three-dimensional (3D) representation of objects and the ability to manipulate, interact, or construct them create deep and effective learning (Hamilton, 2021). Such platforms also allow expanding the scale and reach of learning across institutions and beyond traditional physical classrooms. To date, such environments have not been widely utilized in the field of voice pedagogy and voice science.
Problem: 3D printed vocal tracts allow the visualization of structures that are otherwise invisible during singing/speaking. Interacting with these physical models allows students to understand vocal tract resonances and source-filter theory in a deeper, more intuitive manner. These physical models are derived from MRI scans converted to digital models in CAD, and printed in 3D plastic. Although these plastic models proved very useful, this work highlights the opportunity of making this visualized, hands-on learning experience available to individuals without access to these 3D-printed models. A virtual counterpart of this learning experience promises a customizable, accessible, and fully distributable educational instrument.
Methods: Vocal tract imaging datasets of multiple vowels produced by male and female speakers will be generated using hybrid MRI-CT imaging. Utilizing image segmentation, these datasets will then be converted to 3D models and deployed in an interactive programmable game engine. The VR application developed via this game engine will immerse students in a voice acoustics lab where they can virtually experience the acoustic effects of combining various simulated glottal signals with vocal tracts producing various vowels. Students will be asked to share their qualitative responses to this instructional modality.
Results: Screenshots demonstrating the VR experience and a summary of the students’ qualitative responses after using this educational VR experience will be collected.
Shenandoah University, University of Iowa & University of London Faculty Collaboration
Physical Modelling of High-Resolution Hybrid CT+MRI Vocal Tract Images
Introduction: 3-D printed vocal tract models have multiple applications. In voice science they have been used to explore the changes in vocal tract shape when lowering F0 (Hirai et al., 1993), when singing different pitches (Takemoto, 2012) and when producing different vowels (Kitamura, 2008; Honda, 2010). In voice pedagogy 3-D printed models are a tool for teaching principles of vocal tract resonance and source/filter theory (Meyer, 2019). They are also used in the vocal tract organ – a new musical keyboard instrument (Howard, 2018).
Problem: MRI-based models have limitations: scan resolution and time are correlated, but voice task duration cannot exceed one exhalation (<20 seconds). Spatial resolution for these tasks is often 2mm3, sufficiently accurate only for gross vocal tract structures. MRI uses the magnetic properties of hydrogen molecules found in water and fatty tissues. These hydrogen molecules are only minimally present in bone (e.g. teeth), limiting MRI modelled oral cavity fidelity. Computed tomography excels at visualizing bony structures and tissue-air boundaries, but the risk of ionizing radiation has precluded its use in vocal tract modeling. This risk can be mitigated but not eliminated. If an extremely low-dose CT scan were blended with high-resolution MRI, a highly accurate hybrid model may be constructed.
Methods: Participants will include two professional classical singers (baritone and mezzo-soprano). Each will undergo one ultra-low-dose CT scan (Siemens SOMATOM Force) and multiple high-resolution MRI (GE SIGNA Premier 3T) of six sung vowels. Scans will employ algorithms based on sparse sampling and constrained reconstruction, and a novel accelerated model based deep learning reconstruction approach to enable a high spatial resolution (1mm3) and a short scan time (10 secs). CT and MRI scans will be blended to include bony structures in the vocal tract, and clearly delineated tissue-air boundaries. The blended images will be exported to CAD files and printed in 3D.
Results: Models will have more accurate spatial resolution and acoustic fidelity than has previously been achieved. CAD files will be imported into I-VR and shared via a creative commons license. Implications for voice science, voice pedagogy, and for the Vocal Tract Organ will be discussed.
Shenandoah University Student & Faculty Collaboration
Minimal CT Dosages Necessary for Blended CT+MRI Vocal Tract Imaging
Introduction: Vocal tract models derived from magnetic resonance imaging have inherent inaccuracies. MRI uses the magnetic properties of hydrogen nuclei that are abundant in water and fatty tissues. It cannot accurately visualize bony structures (e.g., teeth), as bone contains very little hydrogen. The oral cavity makes up approximately half of the vocal tract and it cannot be accurately modelled with MRI alone.
Computed tomography (CT) is excellent at visualizing bony structures and air-tissue boundaries. If an extremely low-dose CT scan of the vocal tract were blended with high-resolution MRI scans of the soft tissues, a hybrid CT-MRI vocal tract model could be generated. Advances in medical imaging have led to substantial CT dose reductions, but the risk of ionizing radiation inherent in CT remains, and must be mitigated. Determining minimum CT dosages for blended CT-MRI vocal tract model generation is a necessary first step, and the focus of this study.
Methods: This exploratory study used a donor cadaver from the University of Iowa Deeded Body Program to compare CT (Siemens SOMATOM Force) image quality at various quality reference mAs (QRM) when using dose modulation with 120kV. As the QRM lowers, the radiation exposure (measured in milliSieverts/mSv) decreases. By comparison, a single chest x-ray is about 0.1 mSv. The minimum dosage that maintained clear bone contours and air-tissue boundaries was determined.
Results:
- Neck QRM-10: 0.17 mSv
- Neck QRM-5: 0.09 mSv
- Head/Neck QRM-5: 0.06 mSv
The lowest dose (QRM 5) produced images that allowed clear visualization of the bone and air-tissue boundaries. An exposure of 0.06 mSv represents a substantially mitigated risk for living participants in future studies. Emerging photon counting CT technologies may permit further reductions in ionized radiation, higher spatial resolution and reduced image artifacts.
Shenandoah University Student & Faculty Collaboration
Formant/Harmonic Coupling Patterns in Contemporary Belt as Observed in “And I Am Telling You” from ‘Dreamgirls’ as sung by Amber Riley
Acoustic analysis of the singing voice seeks to quantify the singer’s radiated sound spectrum by examining the strength of the harmonics and their interaction with vocal tract formants. Typical formant/harmonic couplings associated with contemporary singing are: Belt = F1/H2 (fR1/2fo), Mix = F1/H1-H2 (fR1/fo-2fo), and Legit = F1/H1 (fR1/fo) (Titze, 2018). While much of belting utilizes an F1/H2 coupling, belt may also have a strong harmonic boost between F2/H3-H4 (fR2/3fo-4fo) (Dargin 2017; Flynn, 2020). Other formant/harmonic coupling patterns may be present in belt, particularly above F1. This study will explore these additional formant/harmonic coupling patterns.
Objective: This non-experimental, associational study will explore the formant/harmonic and formant/formant relationships for belt notes in the contemporary musical theatre song “And I Am Telling You” from Dreamgirls, sung by Amber Riley.
Research questions:
- If an F1/H2 (fR1/2fo) coupling is observed, what formant/harmonic relationships exist with F2 (fR2), F3 (fR3), and F4 (fR4)?
- If an F2/H3-H4 (fR2/3fo-4fo) coupling is observed, what formant/harmonic relationships exist with F1(fR1), F3 (fR3), and F4 (fR4)?
Methods: Sustained belt notes with a duration ≥0.5 seconds will be extracted and analyzed with Praat 6.1.42 to estimate fundamental frequency and formants 1-4. Formant data will be superimposed onto the LTAS power spectrum for each sample in Voce Vista Video Pro 5.4.2 to examine formant-harmonic couplings. Resulting data will be recorded in JASP, an open-source program for statistical analysis. Audio samples will be taken from a 44.1kHz, 16-bit recording. The voice and instrumental accompaniment will be isolated prior to acoustic analysis by a professional audio engineer using Rebalance by iZotope. Acoustic analysis will be performed on audio samples pre- and post- iZotope isolation to examine the spectral effects of this processing.
Results: Results from this study may guide future research and inform teaching practices.
Shenandoah University Student & Faculty Collaboration
Older African American Singers and Singing Cessation
Objective: The purpose of this study was to investigate the reason and meaning of singing for elderly African-Americans who have stopped singing and how cessation affected their lives.
Research Questions:
- What does being a singer and singing mean to this population (self-definition and social community position)?
- What were the factors for cessation (vocal and non-vocal reasons)?
- How did the participant mitigate cessation (QOL questions are folded into this)?
- What is the participants’ view on cessation (is it permanent, reversible etc.)?
The older African American population has grown in numbers. Longer life spans require medical and QOL solutions to promote good physical health and social well-being. A lack of research has left this growing sector under-served with potential benefits unrealized. Singing as a participatory activity has been found to help improve QOL and thereby possibly promote improved health outcomes and longevity.
Methods/Design: The design is an exploratory qualitative, comparative study model using case studies. The work was facilitated by open-ended interviews in a three-stage process over a period of several weeks. Six participants were enrolled and ages ranged from 50 to 102 years old. The projected conclusion of the study is by March 2022.
Results are pending analysis
The program also presented two papers at the International Physiology & Acoustics of Singing (PAS7+) Conference in Vienna, Austria, last month. Shenandoah University Associate Professor and Director of Speech-Language Pathology Katrina Miller, Ed.D., is presenting with one of Shenandoah’s graduate students at the American Speech-Language-Hearing Association (ASHA) 2022 national convention in New Orleans, Louisiana, in November.