Rob and I built ShoeVox over the past few weekends after being intrigued by the possibilities of Kinect-style voice control for PC applications.
We're using the speech recognition engine built into the .NET framework (presumably based on technology from Microsoft's TellMe acquisition) to detect a fixed grammar. The set of programs and commands for each are configurable by a simple XML file.
We're also leveraging WMI to monitor OS process creation/deletion events, and only run the SRE when a known target program is running to save CPU resources.
When Microsoft releases the Kinect SDK, we'd like to take advantage of its microphone array to do better background noise filtering. We also want to expand support to include browser plugins, as so many people now use Hulu and Netflix as their primary means of media consumption; if you have ideas about how to automate browser plugin UIs, we'd love to hear from you.
I was pleasantly surprised by how easy and high-quality speech recognition was with .NET, and I'd encourage others to experiment with it. We have some other ideas for how to leverage speech in desktop applications, but suffice it to say there is a raft of opportunity in the space beyond traditional dictation software.
Feel free to ask any questions about the speech recognition, process monitoring, keyboard simulation, etc.
We're using the speech recognition engine built into the .NET framework (presumably based on technology from Microsoft's TellMe acquisition) to detect a fixed grammar. The set of programs and commands for each are configurable by a simple XML file.
We're also leveraging WMI to monitor OS process creation/deletion events, and only run the SRE when a known target program is running to save CPU resources.
When Microsoft releases the Kinect SDK, we'd like to take advantage of its microphone array to do better background noise filtering. We also want to expand support to include browser plugins, as so many people now use Hulu and Netflix as their primary means of media consumption; if you have ideas about how to automate browser plugin UIs, we'd love to hear from you.
I was pleasantly surprised by how easy and high-quality speech recognition was with .NET, and I'd encourage others to experiment with it. We have some other ideas for how to leverage speech in desktop applications, but suffice it to say there is a raft of opportunity in the space beyond traditional dictation software.
Feel free to ask any questions about the speech recognition, process monitoring, keyboard simulation, etc.