For just about any task you want to name, a Linux-based desktop can get the job done using apps that rival or exceed those found on other platforms. However, that doesn’t mean it’s always easy to get it to work, and voice recognition is just one of those tricky setups.
A project called Voice2JSON attempts to simplify the use of voice workflows. While it does not provide the actual speech recognition, it does make it easier to get things done and then use speech in a natural way.
The software can integrate with multiple backends to perform offline speech recognition, including Pocketsphinx from CMU, Kaldi from Dan Povey, DeepSpeech 0.9 from Mozilla, and Julius from Kyoto University. However, the code is more than just a wrapper around these tools. The rapid training process produces both a speech recognition tool and an intention recognition tool. So not only do you know that there is a garage door, but you also understand how to open and close the garage door.
Plus, the tools are all designed to work in Unix-style pipelines, which is refreshing. Here is an example of the project website configuration:
[GarageDoor] open the garage door close the garage door [LightState] turn on the living room lamp turn off the living room lamp
There are template features so that you can specify optional words and alternative words in a single rule. There are other features like mapping an object like a living room lamp into something more computer friendly.
Overall, this looks like a fun tool to have in your kit. If you do something interesting with it, be sure to tip us so we can cover it. Meanwhile, we’ve been watching Linux talk for quite a while. Of course, what we really want are voice commands like the USS Enterprise, and we have to admit that it comes close.