Mycroft on Rpi3 does not save registration configuration after registration on home.mycroft.com

I have booted Mycroft on my RPi3 going thru start up sequence instructions and registering new device with boot sequence registration code on home.mycroft.com.
The mycroft software reports that it has registerd sucessfully and responds to “say” commands.
( until it stops working after 30 sec but thats another issue thread)

However when I shutdown and reboot the Mycroft on the RPi has not saved the registration configuration in .mycroft directories and goes thru the same registration code generation sequence.

I have checked the persmission on the .mycroft directories and the mycroft user is able to write to them.

Any ideas on this.

The v0.8 image is a couple of weeks old?

Currently Mycroft Configuration :
( ?? Is there a print_mycroft_config ) command ?
v0.8
Pi3 with official Pi 3 Power supply
Ethernet connection to Internet
3.5 Speaker
No Mic ( yet )

Are you using 0.8a or 0.8b? 0.8b is suppose to fix this issue

8.0a rel at the moment … upgrading to 8.0b

have upgraded to 0.8b
registration key behaviour now is
… mycroft pi prog prompts to register device on clean boot

  • register on home.mycroft.com
  • No response from mycroft pi prog to ak registration
  • reboot mycroft … does not attempt to re-register
  • however sound " say " does not play thru 3.5 mm speaker

Hey there, @southpacific.

Glad to hear the new image is working for you.

Can you clarify what sound “say” does not play through speaker? Are you not getting any audio at all?

v0.8b
Previously with the same externally powered speaker plugged into pi audio out
i used to get mycroft audio e,g what is the time, what is the news

Looking at the log now … the speaker is detected on startup.when i “say hey mycroft what is the time"
the log prints on screen and comes back with
” Connected speaking to mycroft …
sent

Then the “say” command appears to hang and not return to the command line
There is no /tmp/tts.wav file produced

( i dont have mic yet … wondering if this is related …)

Just to verify, your ctrl+c’ing out of the what you first boot into on the pi?

That should be working. You really shouldn’t need a mic or even a speaker, you just need to swap between say commands and view_log
Could you send us a screen shot or ssh into your pi and copy-paste?

OK will do when bac at the “lab” ;->
I retried this “say” yesterday … it gave the same message . on what time is it etc.
started running gave some messages and did not rtn to cli
but there was a /tmp/tts.wav file produced with the correct time answer
and when i used aplay /tmp/tts.wav it played ok.

The speaker is a genric powered 3.5 mm head jack speaker

One thing i have noticed diff btw 0.8a and 0.8b was that in the boot log there the speaker was recognised ( pcm etc) as a surround sound system , a lot of details about number of channels etc.

I may try it wit another speaker before ( like UE Boom ) before i send u the log.

I noticed before the Rpi/Myc would not produce sound thru a a 3.5 headphone jack into a Sony AV rcvr …

Will run some tests

Can you try updating your Picroft manually and letting me know if that helps?

sudo apt-get update && sudo apt-get install --only-upgrade mycroft-core -y

Thanks for the command …
I did not need not need to run as i
I rebooted the rpi and the mycroft 8v version downloaded the next version ( v 8.2 ? ).
Works ok with the generic powered speaker and
cli say hey mycroft ,
runs ok and produces sound now
( and does not hang and returns to cli )
response time is about 3 secs.
Next step is to get a ps3 eye.

Awesome, glad to hear it. And yeah, 8.2 is the latest.

Let us know what you think about the PS3 eye, I just ordered one for testing myself. Seems like one of our best options.

ps3 eye looking a good choice … cheap and good mic.

however issues on the fourth Q+A … hung on the last play wave ,

2017-02-09 08:05:02,131 - SpeechClient - INFO - Speak: Sorry, I didn’t catch that
Playing WAVE ‘/tmp/tts.wav’ : Signed 16 bit Little Endian, Rate 44100 Hz, Mono

see log below

2017-02-09 08:04:33,208 - mycroft.client.speech.mic - DEBUG - Thinking…
2017-02-09 08:04:33,209 - mycroft.client.speech.mic - DEBUG - Waiting for wake word…
2017-02-09 08:04:33,210 - SpeechClient - INFO - Wakeword Detected: hey mycroft
2017-02-09 08:04:33,321 - requests.packages.urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): api.mycroft.ai
2017-02-09 08:04:34,843 - requests.packages.urllib3.connectionpool - DEBUG - https://api.mycroft.ai:443 “POST /v1/stt?lang=en-US&limit=1 HTTP/1.1” 200 None
2017-02-09 08:04:35,628 - mycroft.client.speech.listener - DEBUG - STT: what’s the share price of ibm
2017-02-09 08:04:35,629 - SpeechClient - INFO - Utterance: [u"what’s the share price of ibm"]
2017-02-09 08:04:38,962 - SpeechClient - INFO - Speak: the share price of ibm : $176.17 (IBM, NYSE, 4:00:00 pm EST, Wednesday, February 8, 2017)
Playing WAVE ‘/tmp/tts.wav’ : Signed 16 bit Little Endian, Rate 44100 Hz, Mono
2017-02-09 08:04:58,682 - mycroft.client.speech.mic - DEBUG - Recording…
2017-02-09 08:04:58,687 - SpeechClient - INFO - Begin Recording…
Playing WAVE ‘/home/pi/wavs/shipbellring.wav’ : Unsigned 8 bit, Rate 22050 Hz, Stereo
2017-02-09 08:05:00,545 - SpeechClient - INFO - End Recording…
2017-02-09 08:05:00,546 - mycroft.client.speech.mic - DEBUG - Thinking…
2017-02-09 08:05:00,547 - SpeechClient - INFO - Wakeword Detected: hey mycroft
2017-02-09 08:05:00,552 - mycroft.client.speech.mic - DEBUG - Waiting for wake word…
2017-02-09 08:05:00,640 - requests.packages.urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): api.mycroft.ai
2017-02-09 08:05:01,928 - requests.packages.urllib3.connectionpool - DEBUG - https://api.mycroft.ai:443 “POST /v1/stt?lang=en-US&limit=1 HTTP/1.1” 200 None
2017-02-09 08:05:02,128 - mycroft.client.speech.listener - ERROR - list index out of range
2017-02-09 08:05:02,129 - mycroft.client.speech.listener - ERROR - Speech Recognition could not understand audio
2017-02-09 08:05:02,131 - SpeechClient - INFO - Speak: Sorry, I didn’t catch that
Playing WAVE ‘/tmp/tts.wav’ : Signed 16 bit Little Endian, Rate 44100 Hz, Mono

What did you say? It looks like it heard something and couldn’t turn it to text.

I didnt say anything , “just hey mycroft” before it cut of so the
speech recoginitiion error
and the "speak - i dont catch that , is correct
Looking on another thread ( best microhone ) from the sysadmin, there is a known problem with audio analog on rpi hanging on analog audio.
Will retry with HDMI Audio output.