After updating to ARC 2024.12.18.00 I have a couple problems.

Most important is that the Wake word in Bing Speech will not trigger any of the scripts I use in this skill. It reports in it terminal that it "Heard Wake Word (1)" but it won't trigger any internal scripts. Not any scripts through the attached scripts to the Phrases, not anything in the Start Listening Script Section, not the Script For Not Handled section. However when I switch over to using Auto Record using VAD all scripts are triggered properly with any recognized or not handled phrase.

Next I noticed that when audio is playing through the PC speaker (I don't play audio through a EZB speaker) from one of my ARC PC sound boards this file will be briefly interrupted for a split second when I click on anything in the ARC interphase.  This sucks if the robot is talking and I also need to be in ARC working something.

I'm unable to reproduce those results Dave. The scripts trigger correctly, and there's no pause or anything when I play audio and click on anything in the interface.

After the console logs "heard wake word", if you wait it will send the data that it heard. The "heard wake word" will change the graph to green as it is recording (listening). None of that functionality has changed. 

I'd recommend rebooting your computer. Make sure no other programs are running in the background. Anything that's also trying to use the CPU could be causing issues.

@Dave - thanks for your feedback. Although we cannot reproduce the issue you are experiencing with Bing Speech Recognition, we're curious about the audio pauses. Can you provide more information about how you are playing audio? Are you using scripts or robot skills, or is it text-to-speech? What is the source of the audio? What we mean by that is, are you playing the audio using Text-to-speech or an mp3/wav file? Any information you can give us will be helpful as there are several ways to play "audio," we can't test your experience without knowing how to duplicate it.

Hmmm okay, i'm interested in the answer as well. The audio enhancement was only applied to the audio text to speech engine built-into arc. So I'd imagine there was always a UI performance issue with audio in the past but no one noticed it until it was mentioned and fixed for speech. How about we apply the performance fix to the entire audio library so it works with everything. I'm imagining Dave uses the soundboard (PC) is what he's secretly referring to.

@Dave I'm curious as to what robot skills or code you're using to experience it - if it aligns with what I suggested, it's always been there but I think it can be fixed if the audio fix from yesterday's release is applied to the entire audio library.

We have tested a fix for the symptom that @Dave described with the audio output. While we can't reproduce it in our test environment, we're confident it will resolve the issue on his computer. We will publish an update this evening.

Sorry for the delayed response. Lots going on.

Thanks for the quick attention and help. I'll try to explain better.

Firstly, I've turned off the robot and it's computer overnight. ON powerup today nothing has changed. I still have the two issues I mentioned above.

I'm playing mp3 files stored in PC soundboards in ARC played through the headphone jack of my PC. They are triggered through scripts either residing in the Bing Speech skill or scripts that are in script managers triggered by either Bing Speech skill or the Personality Generator. After more testing I see that the audio playing pauses for a split second when something on the top menu is selected. Something like changing vestural desktops of saving the project. I can't detect any pause when I click on a start button to start a script or other skills like the Auto Position skill. So it looks like it's confined to just top menu clicks. If this has always been there I have never noticed it happening. It's not deal killing issue and really doesn't effect my robot animations unless I'm actually in working in ARC.

The more worrisome is the wake work in Bing Speech not working. @DJ says once the wake word is recognized the display should turn green and start listening. I see the red line graph bump up when I say the wake work (the console logs "heard wake word"). Then as I give a command I see the line graph keep bumping up but it stays red. Then when I switch over to VAD and I speak, the line graph turns green right away and triggers scripts properly.

If it matters, I also need to say that the control computer that has ARC running is on board my robot and connected to three EZBs through the USB ports. I'm not connecting over WIFI.

I hope this helps.

Dave, thanks for your response. Please answer these questions about the Bing Speech recognition...

1. Do you hear the wake word sound when the text "heard wake word" is displayed?

2. Have you changed the wake word sound from the default value to something custom?

3. Does the robot have an internet connection? Bing speech recognition requires an internet connection

4. Is there any other text displayed in the log window? Is the only text you see is "Heard wake word"? Are you confident there is no other text in the log window?

These details are essential to understanding the experience; anything you can provide is vital to solving your problem.

A ha, Looks like you nailed it. Thanks!! I did have a custom sound file for the wake word sound. I replaced it with a different one and everything started working. I haven't had a chance to thoroughly test it but Bing's wake work function seems to be working now.

OK, here's what I was using for a sound file for the wake word sound before I replaced the file. I do not want to hear a sound when the wake word is heard. I have visual lights and movement on the robot that signals the wake word has been heard (triggered in a script in the Start Listening Script Section). So since there is no way to disable this wake word sound in the Bing Speech skill I created a very short silent sound file with Audacity and pointed to that for the wake word (no)sound. It's worked nicely for a long time till this update. Do you have any advice on what needs to be done now to make a silent sound file that will work here?

I had asked a while back for a future update that would give us the ability to disable the wake word sound as there is no way to do do that. It sounded like you all could do that and would get around to it when able. Is that still on your to do list? love