Welcome to the website of our research project ProVoiceVR!

Here you will find information on the background, process and purpose of the project as well as the latest news.

September 2024

Ende September waren wir auf der 12. Tagung der European Speech and Language Therapy Association (ESLA) vertreten, um unser Projekt dort vorzustellen und internationale Kontakte zu knüpfen. Hier finden Sie unsere Präsentation zum Download.


Juni 2024

Auf dem 52. dbl-Kongress in Oberhausen konnten wir sowohl unser Projekt im Rahmen eines Postervortrags präsentieren als auch die VR-Brille am Stand der Hochschule Osnabrück von Interessierten ausprobieren lassen. Die positive Resonanz hat uns sehr gefreut und wir sind sehr gespannt auf die Ergebnisse der zweiten Erprobungsphase, die aktuell mit je acht Therapeut*innen und Patient*innen anläuft.

Unser wissenschaftliches Poster über die Ergebnisse der ersten Erprobungsphase können Sie hier herunterladen.


September 2023

Die erste Erprobungsphase für unsere Anwendung ist abgeschlossen! Aus 20 ausführlichen Interviews mit Nutzer*innen unserer Anwendung konnten wir zentrale Arbeitsaufträge ableiten und befinden uns nun in der Entwicklungsphase, um die Software zu verbessern und neue Funktionen einzubauen.

Bis zur nächsten Erprobung, die Mitte 2024 beginnen wird, möchten wir unter anderem folgende Features in die Anwendung implementieren:

  • Grafisches Biofeedback für bspw. Lautstärke und Tonhöhe
  • Virtuelle Kindercharaktere sowie eine Kindergarten-Umgebung
  • Anpassbare Hintergrundgeräusche (Lärm und Gespräche)
  • Live-Übertragung der Simulation auf das Tablet

Wir danken allen teilnehmenden Patient*innen und Stimmtherapeut*innen für ihr umfangreiches Feedback und sind zuversichtlich, auf dieser Grundlage die ProVoiceVR-Anwendung deutlich verbessern und für die Nutzung in der logopädischen Praxis optimieren zu können!

Im Rahmen unseres Forschungsprojekts ProVoiceVR bauen wir eine Datenbank mit einer Vielzahl verschiedener Stimmbeispiele auf. Anhand dieser möchten wir ein Computerprogramm mithilfe maschinellen Lernens darauf trainieren, die Stimmqualität einer Audioaufnahme einschätzen zu können. Unser Ziel ist es, das fertige Programm in unsere VR-Anwendung ProVoice zu integrieren, um Patient*innen z.B. im Heimtraining automatisiertes Feedback über ihren Stimmklang anbieten zu können, um die Inhalte der logopädischen Stimmtherapie zusätzlich zu vertiefen.

Um dieses Vorhaben verwirklichen zu können, benötigen wir die Mithilfe vieler Freiwilliger, die ihre Stimme dafür zur Verfügung stellen. Das Anfertigen einer Aufnahme für die Datenbank dauert nur wenige Minuten. Dabei erheben wir keine rückverfolgbaren Daten zu Ihrer Person, sodass Sie dabei anonym bleiben. Konkret nehmen wir einmal einen gehaltenen Vokal („aaah“) auf und danach eine neutrale Kurzgeschichte, die von Ihnen vorgelesen wird, wobei es keine Rolle spielt, wie Ihre Stimme klingt.

Nach Abschluss des Projekts möchten wir die Datenbank auch anderen Forschungsteams für wissenschaftliche Zwecke zur Verfügung stellen.

Im Rahmen unseres Forschungsprojekts ProVoiceVR bauen wir eine Datenbank mit einer Vielzahl verschiedener Stimmbeispiele auf. Anhand dieser möchten wir ein Computerprogramm mithilfe maschinellen Lernens darauf trainieren, die Stimmqualität einer Audioaufnahme einschätzen zu können. Unser Ziel ist es, das fertige Programm in unsere VR-Anwendung ProVoice zu integrieren, um Patient*innen z.B. im Heimtraining automatisiertes Feedback über ihren Stimmklang anbieten zu können, um die Inhalte der logopädischen Stimmtherapie zusätzlich zu vertiefen.

Um dieses Vorhaben verwirklichen zu können, benötigen wir die Mithilfe vieler Freiwilliger, die ihre Stimme dafür zur Verfügung stellen. Das Anfertigen einer Aufnahme für die Datenbank dauert nur wenige Minuten. Dabei erheben wir keine rückverfolgbaren Daten zu Ihrer Person, sodass Sie dabei anonym bleiben. Konkret nehmen wir einmal einen gehaltenen Vokal („aaah“) auf und danach eine neutrale Kurzgeschichte, die von Ihnen vorgelesen wird, wobei es keine Rolle spielt, wie Ihre Stimme klingt.

Nach Abschluss des Projekts möchten wir die Datenbank auch anderen Forschungsteams für wissenschaftliche Zwecke zur Verfügung stellen.

Our goal is to develop a virtual reality (VR) therapy application for speech-language patholog and other voice therapistswork with people who are professionally exposed to a high degree of vocal stress.

Unfortunately, professional speakers are predestined to develop voice problems and disorders due to the high vocal effort required on a daily basis. Although voice therapy can create the conditions for healthy and careful vocalisation, transferring what has been learnt into everyday working life is always a hurdle that requires a lot of time and discipline from those affected.

This is why we are developing ProVoiceVR a tool that makes it possible to simulate everyday working life within a protected therapy environment in order to facilitate this kind of transfer in an efficient way.


To do this, we design a variety of virtual environments in which patients can experience situations from their everyday work through virtual reality, allowing them to practise the voice and speech techniques they have learnt in therapy under more realistic conditions. This is supported by various feedback functions.

 The scientific planning, realisation and evaluation take place at the Osnabrück University of Applied Sciences , while the Osnabrück based start-up company VReedback  is responsible for the technical implementation.

The project is financially supported by the German Federal Ministry of Education and Research. gefördert.

Background

Around 6-15% of the total population are affected by a voice disorder, with professional speakers being at a significantly higher risk. Female teachers have a particularly high prevalence of up to 50%. Voice disorders can be characterised by rapid vocal fatigue, a hoarse, rough or breathy voice, a sore throat and even complete loss of voice (aphonia). Voice disorders are associated with high individual and societal costs, including the inability to work.

The aim of voice therapy is to enable patients to control direct (vocal and speaking technique) and indirect factors (stressors or environmental conditions) by themselves in order to achieve a healthy voice.

In particular, the transfer of vocalisation learning experiences from the therapy setting to everyday situations often poses a central problem. This is mainly due to the complexity of everyday situations (e.g. classroom, gym) with variable and less predictable conditions such as background noise or listener reactions, which require a flexible response in terms of voice and speech compared to the therapy situation in a more manageable individual setting.

The simulation of specific speaking situations using virtual reality (VR) could provide a solution to this problem, as the learning experiences during therapy can be partially transferred to the real situation. Studies have already shown that a person's vocalisation in a VR speech situation (in virtuo) is comparable to a real speech situation (in vivo). In contrast, however, vocalisation in vivo and in virtuo differs significantly from a performance practice situation in a one-to-one setting.

The simulation of speaking conditions using VR can enable voice therapy patients to train in everyday vocally stressful situations under supervision from therapists. Transportable VR equipment is also suitable for additional home-based self-training. This gives people with voice disorders the opportunity to undergo high-frequency training and possibly achieve faster and more sustainable therapeutic success. The use of VR has already yielded positive rehabilitation results in other therapeutic contexts such as physiotherapy and psychotherapy.

This research project aims to develop the VR application ProVoiceVR for the use of VR in voice therapy. To achieve this, the expertise of the start-up VReedback GmbH & Co. KG in the development of VR training software and machine learning is combined with the voice therapy expertise, field access and methodological competence of the Department of Speech-Language Pathology at Osnabrück University of Applied Sciences. ProVoiceVR is intended to enable users to specifically train voice therapy-relevant and realistically simulated situations. The user can receive visual feedback on speech behavior (e.g. speaking pace, volume, pitch) and voice quality using AI-supported methods, while the feedback options and training parameters (e.g. adjusting background noise) can be configured by the voice therapist in advance, but also in live operation.

Development process

Our research project is scheduled to be completed in fall 2025. In order to achieve the project goals by then, we are specifically involving voice therapists and patients in the development in two trial phases. For this purpose, around seven therapists will each receive a pair of VR headset with our application prototype and try it out with 1-2 patients each in several treatment sessions. We then conduct qualitative individual interviews with all participants about their experiences with the application, evaluate them and further develop the application based on this feedback.

With this, we want to create the basis for a practical VR application that can be optimized and scientifically evaluated in further research projects after this project has been completed. The aim of this small-scale, user-centered design process is, of course, to be able to make the finished application available to speech-language pathology practices and other voice therapy facilities.

Gallery

Team

Marcel Herrmann

Sales Manager at VReedback 

  • Training of participants
  • Development of virtual scenarios

Tobias Plock

CEO at VReedback 

  • Front-end development (user interface, app design)
  • Web design, technical setup of the voice database
  • Florian Herrmann

    Technical Manager at VReedback

  • Back-end development (programming of virtual scenarios and feedback functions)
  • AI development
  • Melvin Stubbe

    Project Manager at VReedback 

  • Coordination of the research project
  • Training of participants
  • Front-end development (interior scenario design, animations, character models)
  • AI development
  • Jannis Hansa

    Research Associate at Osnabrück University of Applied Sciences Osnabrück

  • Content structuring and networking in the project, contact person for participants
  • Preparation, execution and evaluation of qualitative user interviews
  • Voice therapy expertise for development of the VR application
  • Planning and conducting systematic literature analyses
  • Prof. Dr. Hilke Hansen

    Project Manager at Osnabrück University of Applied Sciences

  • Coordination of the research project
  • Preparation and evaluation of qualitative user interviews
  • Did we catch your interest and would you like to receive further information?

    Then sign up now:

    Did we catch your interest and would you like to receive further information?

    Don’t hesitate to contact us! kontakt@provoicevr.de

    FAQ: 

    Here you will find the most frequently asked questions.

    Is it possible to participate in the development of the prototype?

    Unfortunately not. We have already found enough cooperating practices and therapy facilities to test the VR application.

    Can the application also be used for other types of disorders than voice-related ones?

    We are developing ProVoiceVR with a focus on voice therapy, but of course the application can be used or tried out for anything where it seems helpful.

    Are the VR goggles supposed to replace therapy?

    No. We are developing the VR application as a tool and aid for voice therapists to make everyday situations and transfer in general easier to train. It should and can only be used as a supplement to voice therapy treatment.

    When will it be possible to buy the application?

    The current development process is expected to be completed by the end of 2025. Depending on the remaining development work, our aim is to start sales soon after.

    Exact launch dates will be announced on this website.

    How much will the application cost?

    We can't give an exact answer at this stage. Costs will arise for the required hardware (VR headset and tablet), our developed software (ProVoiceVR) and practical training on how to use it. The costs may be reduced if compatible hardware is already available.

    Is the weight of the headset a problem?

    We have occasionally received feedback that the weight could become too much in the long term, especially for patients who have difficulties with their head posture and/or cervical spine. However, the weight was not a problem for many participants.
    The new generation of VR headset used represents a major step forward in terms of wearing comfort, among other things.


    It may also be possible to prevent this problem by shortening the exercise sequences.

    Which VR headset model do you use in the project?

    We use the Meta® Quest 2 and Quest 3 models.

    Self-awareness is particularly important in voice therapy. How does this fit in with the focus on biofeedback?

    Biofeedback is just one function of the application that can be switched on and configured individually. If its use is not considered beneficial, or if, for example, patients explicitly have to practice self-assessment and self-awareness, it can simply be deactivated.

    In our view, the biofeedback functions are most important for possible use in home training - of course after prior consultation during therapy, but then without direct therapeutic feedback. Even the best biofeedback cannot replace the latter, but ideally it can be used to train individual aspects of voice use, such as volume, pitch or speaking pace.

    en_USEnglish