Lol Frank mine often accidentally hears and speaks to its self...nothing makes you feel a proud parent of an intelligent artificial life form like catching it in an endless Hello. Hello. loop.

So it’s very doable but I try to design against it happening.

I would consider all of your variables. First is mic, I have rather good quality Blue Yeti desktop mic that easily picks up all sounds.

Next is speakers. My robot usually speaks through PC speakers which are going to be much clearer than the EZB speaker, though my mic can hear that fine as well.

Then you have Windows settings as you mentioned.

Then you have ARC speech accuracy you can adjust.

Then you overal volume level as too high or too low might distort.

Then the actual voice, as you mentioned.

Hey there Frank! I have the same thing happening with Dorian quite often where he hears himself and tries to respond to that, and just as Justin I usually script against that happening. I can certainly see why you want something like this to work though! I also intend to have robots talking to each other later on.

I'd say the microphone is very important for this to work, I'm also using the Blue Yeti microphone and Dorian hears himself despite currently using the tiny EZB speaker (in the process of changing it to something bigger now) and the sound quality being really terrible since it's mounted in a strange way inside his chest. So with the right microphone you can definitely get this to work.

@JustinRatliff and @Sebediah,

Thank you very much for the replies. I just ordered the Blue Yeti mic.

What still surprises me is that JD could here me very well. but if I recorded my voice and played that back, JD did not hear it

Regards, Frank

Hmm that is a bit strange indeed. What did you use to play the recording of yourself? If it was done through the EZB speaker or a bad laptop speaker then it would be more difficult for speech recognition to pick it up compared to your own voice.

You could always cheat and do an MQTT back channel so it appears as if they are talking. I guess it is a bit like trying to talk when you can use mental telempathy (but it is polite to keep the words going in the foreground so others can listen in).

@Sebediah I recorded myself on my iPhone 7 Plus using MP3 recorder IOS app. I then played it back on the iPhone. Its funny, it liked my voice as I was recording it, but not on playback

To my ears it sounds the same, but not to JD's ears, at least until I get my new microphone. I was hoping to use Sound Board (EZB) if it liked my recorded voice to remove the TTS issues

@Nink, I was able to fake the conversations very nicely by connecting to two EZBs in once project. Really solves any timing issues as well as speech recognition issues

Regards, Frank

Yeah I suppose even when it sounds good to us, the microphone might not agree, haha. Anyway, as stated I'm confident the new microphone will do the trick for you as it has worked well for both me and Justin. It's a shame really since at some point this winter I'm going to have to start looking for something else to use with my robot, as I intend to bring him with me to places later and I want something smaller and less clunky out there than the Yeti, but it sure is a good microphone.