If your thumbs are tired of typing all your texts and emails, there's a new keyboard that lets your voice do the work, and it won't leave you worrying about errors.
TalkType is a free third-party Android keyboard from the Chinese web services company Baidu that replaces the QWERTY key interface with a big microphone button, which it claims is more accurate than actual typing.
To prove that speech recognition is indeed superior, Baidu teamed up with Stanford University for a study that tested the firm's Deep Speech 2 software, which beat humans at typing faster than you can say Deep Blue. The study, performed with 32 texters between 19 and 32 years old, showed that Baidu's tech is three times faster than typing in English, with an error rate 20.4 percent lower than someone hunting and pecking on the screen.
MORE: 10 Best Messaging Apps
We tested Baidu TalkType with a couple sentences, and found it to be a clear winner over the voice recognition in the Galaxy Note 7's Samsung keyboard app. TalkType picked up Echinacea, a word we tossed in to trick it, which Samsung's voice recognition did not. TalkType leads to less editing, as I only needed to add one word to its text, while Samsung's errors made me delete three words and add two.
The Standord study consisted of subjects typing or speaking "about 100 phrases sourced for a standard library of everyday phrases used in text-based research," such as "physics and chemistry are hard" and "go out for some pizza and beer." And for those moments when speech recognition isn't enough, say you're using a complex word that is hard for the software to pick up, TalkType includes a standard keyboard for manual entry.
Half of the subjects performed the task in English using the iPhone's stock keyboard, while the other half conducted the test in their native Mandarin using iOS’ Pinyin keyboard. Mandarin typists saw an even larger increase in accuracy, with their error rate dropping 63.4 percent when speaking.
“We knew speech recognition is pretty good, so we expected it to be faster," said study co-author Sherry Ruan, a computer science PhD student at Stanford who helped run the experiments. Ruan and her colleagues "were actually quite surprised” by the results.
Though some may want to hold onto their traditional keyboards forever, Stanford computer science professor and study co-author James Landay is bullish on speech recognition, saying “We should put speech in more applications than just typing an email or text message.”