I’ve tried to use speech recognition for many years. Tried a variety of solutions that run on Windows PCs and none of them have really performed very well, even after extensive training. But now I'm using speech recognition from Google and it's working very well.
What do I want from speech recognition? To be able to speak naturally, and have a minimal set of updates afterwards.
The Google Speech API is as close as anything that I’ve used. I use it from Google Docs with an add-on from their collection. The add-on I chose was Speech Recognition from efv-solutions.com.
When the add-on is loaded a new pane shows up on the right side of a Google document. Press the start button and start talking and your text shows up in the document. There's a delay of a few seconds while backend services process your words before it shows up so it's a little disconcerting to try to watch it. It's best not to try to watch, just talk.
Why am I telling you this? It's because speech recognition turns out to be a big data problem. To understand speech takes context and Google has that context in this fast database. The standalone PC applications had no context and failed rather amusingly sometimes but mostly just badly.
Context is why I think our customers succeed with Zenoss. Our Live Model gives a business context to the rather vast amount of raw data we collect, helping teams rapidly solve service problems.
I just wish that Google knew how to spell Zenoss. Genius. Cinos. Venus. Mouse. He knows. Casinos. Xenos. Ninos. Dino’s.
Of note - This document was typed with speech recognition.
Interested in learning more? Enter your email address below to subscribe to our blog!