Turn-key PCB assembly services in prototype quantities or low-volume to mid-volume production runs

Speech recognition by bare Arduino

You probably think that speech recognitions is very intense task for processors. In fact it is so if you expect great accuracy and flexibility. But in some cases you simply need electronics to understand few voice commands and perform them with decent accuracy. This is where a μSpeech Arduino library may help. The library is pretty simple and effective. It avoids using FFT but uses voice fricatives to differentiate words. By using simple math to determine words or letters it is able to recognize preselected phrases with up to 80% accuracy.

Well you cannot expect much more with Arduino which is very limited in processing power and especially in RAM. Speaking of efficiency – it uses about 160 bytes of RAM to process and takes around 3.2 milliseconds to recognize. So plenty space and power to do other desired tasks controlled by voice. The library is constantly improved that may give even better results. SO this is a great choice for simple voice enabled controls with Arduino. By the way hardware is very simple – single transistor amplified electret microphone attached to Arduino analog pin.

[..Source link..]

Bookmark the permalink.

Comments are closed.