As a personal project, I made a diary app using speech recognition. There are many hand interaction-based diaries: either via typing or by physically writing with a digital pen. However, there are no apps dedicated to writing diaries solely by voice. LaLa lets users use speech as prime means in writing diaries, and provides a visually structured interface to keep the diaries organized.
LaLa is a mobile diary app you write with your voice. Instead of having to type in the texts or write with a digital pen (e.g. Apple Pencil), you simply speak to the app to write your diary.
LaLa is a mobile diary app you write with your voice. Instead of having to type in the texts or write with a digital pen (e.g. Apple Pencil), you simply speak to the app to write your diary.
Speech-to-text interactions have become readily accessible for both developers and end-users, but it still falls short of human-level dictation. Furthermore, there still remain critical challenges to user experience that cannot be solely handled with STT gadget, such as detecting and writing down the punctuation marks from human speech.
One of the technical challenges in implementing speech recognition for an app was handling failures caused by resource limitations. Due to iOS resource limitations, by default, each session of a speaker instance can stay alive for only a minute at max. Because users can and most likely will sometimes leave the recording session on for more than a minute, it was important to take the system-level limitation into consideration to guarantee continuous and intuitive user experience.
This project was a way to actualize the idea of leveraging speech interactions to explore users' private thoughts. Current speech interactions mostly involve looking up facts. However, I believe that speech interactions can also benefit users as a reflective way to record their thoughts, since when speaking out, users hear their own thoughts through their voice, instead of only keeping them quiet inside their minds.