Paper Reading #3: Voice Typing: A New Speech Interaction Model for Dictation on Touchscreen Devices

Voice Typing: A New Speech Interaction Model for Dictation on Touchscreen Devices

Anuj Kumar, Tim Paek, Bongshin Lee

Microsoft Research

One Microsoft Way

Redmond, WA 98052, USA

{timpaek, bongshin}@microsoft.com

Human-Computer Interaction Institute,

Carnegie Mellon University

Pittsburgh, PA 15213, USA

anujkl@cs.cmu.edu

The researchers introduced a new speech interaction model called Voice Typing. Voice Typing allows users' utterances to be transcribed as they are produced. This creates real-time error identification for in speech recognition software whereas other voice typing software only allows users to check or make corrections after their "speech". 

The users tested the software by in a email composition task. The users were asked to compose an email with a given structure. The email could be filled out by the users themselves. Each user composed 2 to 3 practice emails for each experimental condition (4 experimental conditions). Much of the data gathered was statistical. This included data such as number of substations, deletions, insertions made and how much was uncorrected. Users were also asked to rank the software qualitatively. 18 out of 24 users preferred Voice Typing over Dictation.

This isn't anything new, but that doesn't mean it wasn't interesting. The pros and cons in voice typing and dictation are interesting in that one can outweigh the other depending on the situation or problem. There are many other Voice Typing software programs out there, but there's also lots of room for improvement.