Response Time Error Handling

Hello, is there a way to abort or disregard the Call Sign or a trained sentence if the Recognizer Beep takes longer than X milliseconds (after speaking the Call Sign or sentence?

I ask because I’ve developed an eating and drinking machine for quadriplegics who may have reduced lung capacity and therefore lower dB output when speaking. I’ve found out that if user speaks Call Sign or a trained sentence and the beep is delayed more than approx 1 second there’s a decent chance that the wrong sentence # will be recognized.

Note- I’ve set Threshold to 3 and am using an ETM-008 directional mic with a sensitivity of -38dB.

thanks!
mg

I am not yet completely understanding the question: Why would the beep take longer? I think the easy solution to this is to slightly increase the threshold. One way to tune the threshold is to use MICDEBUG which will replay whatever the system heard. We want to make sure that noise isn’t recognized as speech.