New to Raspberry Pi - I want to create a device which upon hearing a particular word it will activate/deactivate something. E.g. if I say the word "light", it will turn on/off an LED light. The device will only recognize this one and only word that I program it with. Hoping to use it at home which might have some background noise.

The key part that I'm not sure how to implement is the voice recognition. Would open-source libraries like mycroft and jasper help with this, or is there any other solution? If I were to use it in a home setting with various background noises is rnnnoise the state-of-the-art noise cancelling technique or are the open-source packages enough?

Edit: Sorry I should've specified, I want this device to be disconnected from the internet, so using Alexa / Google would be off the table.

  • how will the device know that you are giving it a command? – jsotola Sep 27 '18 at 5:55

I think the next links can help you:



It explains how to implement voice recognition with the help of Google Voice Commands.

About background noises, I have experimented at work the Google solutions, and it's really a big issue... When your commands can be delayed, you can filter the noises like explained here (https://www.techsmith.com/blog/not-late-reduce-audio-noise-recordings-free/) but sure , it's not easy :)

  • You should include the important content in your answer, so that it remains useful even if the links are gone. – Dmitry Grigoryev Sep 27 '18 at 13:41
  • Hello @DmitryGrigoryev, I have a similar project forcasted. As you sugest, I will prepare a full reply so, effectively, if the links are gone, nothing will be lost :) – Llorenç Corbey Oct 3 '18 at 12:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.