Michaela Nachtrab
Ashoka Fellow since 2011   |   Germany

Michaela Nachtrab

VerbaVoice
Michaela Nachtrab envisions a society in which accessibility for hearing disabled and deaf citizens is the norm rather than the exception. This means seamless communication in educational and…
Read more
This description of Michaela Nachtrab's work was prepared when Michaela Nachtrab was elected to the Ashoka Fellowship in 2011.

Introduction

Michaela Nachtrab envisions a society in which accessibility for hearing disabled and deaf citizens is the norm rather than the exception. This means seamless communication in educational and professional environments as well as access to information, events, and all kinds of media. Michaela introduced VerbaVoice as a platform to jumpstart a self-organized community of deaf and hard of hearing that can take an active role in advancing accessibility and their own participation in society.

The New Idea

Over 75 percent of deaf and hearing impaired citizens in Germany do not know sign language and rely on personal voice-to-text transcription support. The high cost and limited availability of this service severely decreases their chances for educational and professional success. At the same time, affordable automatic speech recognition technology will not be able to provide the necessary quality for another twenty years. With VerbaVoice, Michaela is building a two-tier architecture to overcome this problem: She is pioneering a web-based service that dramatically reduces the costs of transcription services and in addition, she is building the communication platform that enables deaf and hard of hearing citizens to take an active role in ending their exclusion. Michaela’s efforts have spread throughout Germany and will expand across Europe in the coming years.

One fundamental building block is Michaela’s market-based solution for pervasive transcription services. With Michaela’s online platform, a hearing impaired person can book and connect to a transcriber whenever necessary. The voice of the speaker is recorded by a laptop or mobile phone, transcribed in real-time by a transcriber working from home and displayed on the screen of the laptop or phone. This system reduces the costs by 35 percent, offering the hearing impaired a largely available and affordable transcription service for professional, educational, and private occasions.

Beyond her core service, Michaela recognizes the potential for broader social application of voice-to-text reporting: Her transcription service can serve as the basis for a second level of empowerment and integration of hearing impaired and deaf citizens. The beneficiary community can flesh out the social value of the service through numerous initiatives. Beneficiaries can crowdfund subtitling of news or movies for the community or the public. Or they can give everyone, regardless of their hearing status, the opportunity to participate in and follow public events online via transcript—for free, or at a minimal cost. The community platform on VerbaVoice also allows the beneficiaries to organize campaigns and to create and publish rankings of companies and universities that integrate transcription tools most effectively.

The Problem

Michaela works to combat a two-fold problem affecting the deaf and hard of hearing community in Germany: (i) even though transcription is an indispensable tool for communication, it is not accessible and cost-effective and (ii) hearing impaired citizens are denied an active role in integrating themselves within the greater population.

300,000 Germans (and 40 million people worldwide) are hearing impaired to a severe degree, resulting in their social and economic exclusion. Their unemployment rate is significantly higher and their average salaries much lower than those of hearing individuals. Currently, there are three ways for a hearing impaired person to communicate verbally: Lip reading, sign language, and transcription. Only 20 percent of meaning in communication can be read from the lips; the remaining 80 percent must be inferred. In addition, only a quarter of all of hearing impaired persons in Germany (80,000 people) understand sign language, which is a figure that is increasing as older adults become hard of hearing yet no longer learn sign language. The prevalence of sign language will continue to decrease as well because more and more deaf children are receiving cochlear implants, which restore auditory perception but leave the child torn between the deaf and hearing communities.

Even though transcription is the most effective solution to these communication challenges faced by the deaf and hard of hearing, it is difficult to access: Only 10 percent of German TV programs have subtitles and very few public events, seminars, lectures, or courses offer transcription services. The availability of personal transcription is even more limited. Personal transcription support requires a professional transcriber to accompany the hearing impaired person to lectures, business meetings, job interviews, and similar occasions. The transcriber then converts spoken communication into written form. The cost of this service is significant for social insurance systems and the individual clients. In Bavaria, for example, the largest state in Germany, there is currently one active transcriber. Transcribers must travel with their clients and are paid by the hour, in addition to travel costs, which makes the service very expensive.

For decades, scientists and engineers have tried to develop fully automatic speech recognition systems. However, to date, no automatic solution is fully usable for the hearing impaired and even market leaders like IBM, Fraunhofer Institutes, or Nuance estimate that a usable system is around twenty years away. The fundamental challenge is providing, at the same time, a solution that can understand every voice, a broad vocabulary, and can do it in real-time.

In addition to financial and technical obstacles, hearing impaired and deaf citizens are denied an active role in solving this predicament for themselves. They are accustomed to being recipients of public support services, but the highly individual provision of support has made it difficult to share these resources broadly within the community. Public awareness of this exclusion problem is also minimal, reinforcing the lack of integration across communities of hearing and non-hearing persons.

The Strategy

Michaela’s strategy is based on a combination of her online service and its potential for social action which produce an empowering architecture for accessibility.

Regarding her market solution, Michaela has combined the right pieces for a new kind of online conferencing service: The voice of the speaker is transmitted to a central location where the transcriber “re-voices” the content word by word for a speech recognition software that is trained specifically for his/her voice. The software transcribes the spoken words into written form. Here, notably, Michaela does not develop the speech-recognition technology herself but can use off-the-shelf products because her customizing and the individual voice training make standard products fast enough. This also makes her independent of specific software products. As technology progresses in the future, Michaela will be able to plug in new software solutions. In this sense, she is more an online convener than a technical service provider.

The text is then transmitted to a display in front of the client on a cell phone or computer, with a minimal time lag. Michaela’s online platform allows the hearing impaired to book online transcribers with only a short notice. She is currently working with twenty transcribers but is collaborating with existing booking agencies that offer transcribers to hearing impaired citizens to prevent any duplication of efforts.

Most of the 35 percent cost reduction that Michaela has already achieved through this platform is the result of the design that enables transcribers to work remotely. Transcribers can accept orders that would previously be impossible to fill because of travel time and other logistical barriers. The online platform allows many transcribers who work part-time or are disabled themselves to work from home. Finally, it simplifies the complex billing system for the transcription service. The revenue of VerbaVoice provides a share of these reimbursements. In 2010, VerbaVoice provided 900 hours of transcription services to approximately fifty customers.

Michaela’s market solution has tremendous potential that goes beyond just providing a service. For example, Michaela is working on building a library of “voice profiles” for universities and companies. With the help of these individual profiles of speakers, for which speakers have to train the software for a couple of hours, the software can transcribe oral communication directly without using a re-voicing transcriber. If most university professors or the important work colleagues of a hearing impaired person had such profiles, transcription based on automatic speech recognition would be available instantly—and at an even lower cost.

This turns a technical challenge into an opportunity for social action: The community of users can now organize campaigns to convince family, friends, colleagues and teachers to create voice profiles and make themselves and their organizations “accessible.” Also, the community could rank the most accessible companies by the number of voice profiles they have created. Michaela plans to extend her service for sign language interpreters (employing online video technology) to cover the group of hearing impaired that does not understand written language.

Independently of how the transcription methods for her market solution works, Michaela’s platforms open numerous ways of taking social action. For example, the platform is set up to help gather volunteers and sponsors who can offer transcription time for free or at a low cost for private occasions (including parties, weddings, or funerals) that hearing impaired could not otherwise attend due to lack of insurance coverage. In addition, the platform can enable crowdfunding for real-time subtitling for news, a movie or a TV show, since it would be too expensive to pay for transcription for these media types individually. Finally, the platform can be used for announcing which public events (lectures or concerts) have already been booked with a transcriber. Any community member can choose to join, either online or in person. Similar to online-mapping platforms showing the accessibility of buildings for wheelchairs, this will greatly encourage the hearing impaired to participate in social life and engage as active, equal citizens. Michaela plans to collaborate with the major associations as well as the most important online hubs for the deaf and hearing impaired.

In addition to her technology and community development, Michaela is active at the policy level. She is working with the Bavarian State Parliament to ensure transcription services for all parliamentary debates. This not only properly implements the UN Convention on the Rights of Persons with Disabilities, but gives her efforts greater public visibility. In the future, Michaela will apply this to other parliaments and extend her lobbying activities.

The Person

Michaela has worked with the hearing impaired since 1993. She studied Deaf Pedagogy and worked in rehabilitation for more than eight years, building and managing several local branches as a regional director for one of the biggest rehabilitation companies for the hearing impaired in Germany. Michaela witnessed firsthand the tragic bottlenecks of transcription for the hearing impaired.

Since 2002 Michaela has worked part-time as a sign language interpreter. She also created an information platform around financing options for communication support for the hearing impaired and founded Nachtrab Social Affairs, an association supporting projects with social impact. As a result of these roles—which involved her as both a project manager and a sign language interpreter—she became dissatisfied with the great gap in care for the deaf and hard of hearing community—the little to no attention devoted to those hard of hearing or deaf citizens who did not know sign language. Michaela began to develop the idea of VerbaVoice as a merger between social and technical solutions and soon became committed to seeing it done.

Michaela previously resolved to complete her master’s thesis only when she found a topic that would enable her thesis to be relevant, applied, and more than just an item on a bookshelf. In 2008 she completed her master’s degree in business administration with a focus on social entrepreneurship. The stakeholder analysis in Michaela’s thesis forms the basis of VerbaVoice’s business plan.

Are you a Fellow? Use the Fellow Directory!

This will help you quickly discover and know how best to connect with the other Ashoka Fellows.