Inspiration
Our exemplar, Paul Kotler, doesn't use his voice to speak; he uses augmentative and alternative communication (AAC). He mentioned that the slow speed of entering text is frustrating for him. That's true for many users of AAC. Slow text entry hinders free and easy communication, making conversations frustrating. The average qwerty typist produces text at speeds of 50 to 80 words per minute, while the average rate of human speech is 160 to 180 words per minute. The average stenographer can produce text at speeds of 200 to 240 words per minute. This solution might not be workable for Paul because his motor planning issues might preclude using all ten fingers at a time, but for people who use AAC and have sufficient fine motor control, steno input could give them something they've never had before: The ability to communicate at realtime conversational speeds using text-to-speech.
How it works
We started with an open source 3D printed steno machine (StenoBoard, $150 - $300), an Android phone, and an open source app called StenoKeyboard. Our developer, Jacob, wrote new code into StenoKeyboard to allow us to produce realtime text-to-speech as we typed with the StenoBoard into any text editor. Our industrial designer, Rocio, prototyped several potential wearable solutions and eventually combined a piece of expertly bent plastic with some heavy duty velcro, hot glue, and a clip from a 25-foot tape measure, which attaches the StenoBoard to the user's belt and allows them to produce realtime text-to-speech while monitoring the text output on their phone.
Challenges I ran into
The StenoBoard is a bit heavy and cumbersome, but was the only solution that would be fully usable by the end of the weekend. We considered several designs that could be implemented in the future, which would be more portable and less awkward to wear, but we eventually settled on using the StenoBoard, because it's readily available and works right out of the box. We had problems with speed and accuracy when the StenoBoard wasn't mounted firmly enough to prevent flexing and bouncing, but we eventually resolved those by reinforcing the plastic chassis in several places. We're also concerned that the speaker on the Android phone is a bit too quiet, so we might want to add a Bluetooth speaker to the wearable rig in the future. We also encountered a few bugs with the base app (StenoKeyboard) that will have to be hammered out before this is a fully stable and reliable solution.
Accomplishments that I'm proud of
Our fashion plate stenographer test model, Stan, found a wearable steno solution (thigh-mounted steno keys on fabric squares) that he thinks is chic enough to use in social situations, and which works just as easily standing and sitting. It's still just a design idea, not yet a functional prototype, but it's got a lot of potential for future iterations of wearable steno solutions. For people who don't mind carrying somewhat bulkier gear, we now have a fully functional conversational-speed text-to-speech solution that costs only $300 plus the price of an Android phone.
What I learned
We thought that building a speech engine into StenoKeyboard might be trivial, but it turned out to take the entire weekend. The result is impressive; now anyone who knows steno can communicate as quickly and easily in text as other people can by speaking with their voices.
What's next for StenoSpeak for Android
We'd like to further develop the sleeker and more low-profile wearable options. We want to expand our free online textbooks and interactive tutorials to make the steno learning process easier and more fun. We want to get the word out to AAC users who would like to communicate more quickly and without the need of a tabletop or other surface.
Log in or sign up for Devpost to join the conversation.