Experiences with the Pi (audio and wake word)

Cool!

A utility to set up audio is definitely one of our short term goals.
Feel free to join us on slack for faster interaction:D
http://mycroft-ai-slack-invite.herokuapp.com/

FireBit,

Are you plugged in to a network cable or wifi and connected to the internet? If so, it looks like you need to “pair” the device. You should visit https://home.mycroft.ai and enter the code that is given when Picroft comes up. Usually it is something like AW2RGL, although it shows up in the on-screen logs like “A. W. 2. R. G. L.”

You should also pull down the latest version of the image, I slipstreamed a couple minor changes that might help you.

I made some minor changes to the image to bump it to 0.8a. This makes the wifi setup work (it was broken in most cases) and might fix a few issues for people running Blue Snoball mics, at least.

There are a few other issues being reported that are more general, particularly an error that we are seeing with SSL3. Looks like it is a known problem with an underlying library. I’m looking for a fix for that, and when it is straightened out the image should update itself automatically when we put out a new release of mycroft-core.

You can download the 0.8a image at the same link: https://rebrand.ly/Picroft-0_8

Also, I’d encourage people to use the wiki on the Github repo ( https://github.com/MycroftAI/enclosure-picroft/wiki ) for posting ideas and the Issues ( https://github.com/MycroftAI/enclosure-picroft/issues ) for reporting problems. That way we won’t lose track of them. :slight_smile:

Steve, thank you for your reply. To answer your question , yes im connected to the internet via network cable and I have connection. I also paired my device on home.mycroft.ai with the code generated.

Just tested the new image and it’s still not working. In fact its worse than 0.8, when I pair the device, it seems to work but after reboot, its asking me to pair the device again. So far 5 reboots and had to pair it 5 times.

On 0.8 I was getting my Pi to freeze after pairing and reboot. I have cloned the repo and re run the scripts and it stopped freezing.

I have setup my sound card and mic according to guide, and it seem to work, but is very temperamental. I’ve managed to use wake word but only few times, and it gave me the time and weather. It only happen 3 or 4 times and cant get it to work. I have noticed that im getting massive lag between saying command and the pi doing anything. For example I said few commands and about 5 min later those commands show on the screen and can hear the answer. could it be that the PI 2 is not fast enough? I have 200mb/s pipe to internet so cant be my internet.

Randomly I hear “sorry i didn’t catch that” even tho i don’t speak anything. I used alsmixer to adjust mic volume and my speakers but not difference.

Also looked around the folders and cant find identity2.json, just identity.json. My question is, should I have identity2.json?

Cheers,
Firebit

1 Like

I have just today downloaded and installed the latest image and I am having the same issue. I am not using a standard mic/speaker config, so I will try that when I get the hardware in a few days.

I am currently testing with a Nuance USB power MIC with a built in speaker. I do hear the audio response from the speaker, but the only response I am getting is “sorry i didn’t catch that” no matter which of the core commands I use.

I am also having to re-register the device with every reboot.

Hi Steve, I totally agree with you. We are in the process of learning and we certainly will for the next couple of years. From my perspective it is perfectly fine to have some limits now and build up on those until we can outgrow them. What makes MyCroft great, is the ability to free us from the big guys like google, amazon and apple and still get smart-whatever systems. So keep going. I have to admit though, I wasn’t at all able to get my jessie-image to actually output any voice, but it understands me and I get the feedback in the log.