3 ms·
I'd be interested to see low-power voice recognition. My PowerShiri demo uses ~12% CPU on a Core2 Duo laptop when it's listening.
by stephengillie 10y ago
I'd be interested to see low-power voice recognition. My PowerShiri demo uses ~12% CPU on a Core2 Duo laptop when it's listening.
- etatoby 10y agoI'm pretty sure most apps use the default OS services, which just send the compressed voice data over to a server, where the actual recognition is performed (and your voice data stored forever, for "research and development" reasons...)
- imaginenore 10y agoSuch a stream would have been detected a long time ago. Do you have any evidence?
- stephengillie 10y agoIt would be trivial to have .NET make UDP calls based on speech recognition input. But what you're describing sounds more like a phone call, where all audio is transmitted, not just a few bits of data. If you have a pcap file showing the outgoing packets, I'd be interested to see it.
- nshm 10y agoYou need to try CMUSphinx http://cmusphinx.sourceforge.net http://cmusphinx.sourceforge.net. In keyword spotting mode it is very lightweight. Python example is here: https://github.com/cmusphinx/pocketsphinx/blob/master/swig/python/test/kws_test.py https://github.com/cmusphinx/pocketsphinx/blob/master/swig/p...