Snips not speaking


#1

I did a new installation on a RPI3 but I’m unable to get audio working. I followed the documentation on Github, I can record audio and play it back as suggested in the troubleshooting section. Using snips-watch -vvv I see that snips reacts to the hotword and even understands my question, but it never answers anything. No output on the speakers.

Here is my /etc/asound.conf

pcm.!default {
        type asym
                playback.pcm {
                        type plug
                        slave.pcm "hw:0,0"
                }
                capture.pcm {
                        type plug
                        slave.pcm "hw:1,0"
                }
}

ctl.!default {
        type hw
        card 1
}

And this is the output when restarting the sound system: https://gist.github.com/tridex/8c7f57b796c632b44d2adc5621a6d47a

For reference here is the output of aplay -l

**** List of PLAYBACK Hardware Devices ****
card 0: ALSA [bcm2835 ALSA], device 0: bcm2835 ALSA [bcm2835 ALSA]
  Subdevices: 8/8
  Subdevice #0: subdevice #0
  Subdevice #1: subdevice #1
  Subdevice #2: subdevice #2
  Subdevice #3: subdevice #3
  Subdevice #4: subdevice #4
  Subdevice #5: subdevice #5
  Subdevice #6: subdevice #6
  Subdevice #7: subdevice #7
card 0: ALSA [bcm2835 ALSA], device 1: bcm2835 ALSA [bcm2835 IEC958/HDMI]
  Subdevices: 1/1
  Subdevice #0: subdevice #0

Any help would be highly appreciated!


#2

Seems I have the same problem? I added the weather and lights bundles to My Assistant on Unzip noticed that I didn’t get the dataset.json file in the assistant directory could this be why it is timing out? See time at _[08:57:09] below.
Not sure what to try next…

[08:56:58] [Dialogue] session with id ‘50b41a30-c566-489e-aa55-f7f6e0fc2570’ was started on site default
[08:56:58] [Asr] was asked to listen on site default
[08:57:03] [Asr] captured text “bedroom light on” in 5.6s
[08:57:03] [Asr] was asked to stop listening on site default
[08:57:03] [Nlu] was asked to parse input bedroom light on
[08:57:03] [Nlu] detected intent lightsTurnOnSet with probability 0.593 for input "bedroom light on"
Slots ->
house_room -> bedroom
[08:57:03] [Dialogue] New intent detected lightsTurnOnSet with probability 0.593
Slots ->
house_room -> bedroom
[08:57:09] [Dialogue] session with id ‘50b41a30-c566-489e-aa55-f7f6e0fc2570’ was ended on site default. The session was ended because one of the component didn’t respond in a timely manner
[08:57:09] [Asr] was asked to stop listening on site default


#3

Me to… Seems snips is working, but she skill donsnt reply. I cant fugure out what is wrong. I have made many freshs installs following the guide on Github, but no matter what snips is only makeing blim and blam, and never say anything


#4

So I had the same issue and couldn’t get any output by default either. So you can test by sending messages directly to the snips audio server or tts service.

First, run snips-watch -vvv

Here is an example of speech input, the audio came from elsewhere though

root@raspberrypi:~# snips-watch -vvv
[23:00:35] Watching on localhost:1883
[00:46:16] [Hotword] detected on site default
[00:46:16] [Asr] was asked to stop listening on site default
[00:46:16] [Hotword] was asked to toggle itself 'off' on site default
[00:46:16] [Dialogue] session with id 'cd6fe08f-e9cc-4c4c-a080-61418a763276' was started on site default
[00:46:16] [Asr] was asked to listen on site default
[00:46:21] [Asr] captured text "play metal playlist" in 4.9s
[00:46:21] [Asr] was asked to stop listening on site default
[00:46:21] [Nlu] was asked to parse input play metal playlist
[00:46:21] [Nlu] detected intent playPlaylist with probability 0.894 for input "play metal playlist"
              Slots ->
                 playlist_name -> metal
[00:46:21] [Dialogue] New intent detected playPlaylist with probability 0.894
              Slots ->
                 playlist_name -> metal
[00:46:26] [Dialogue] session with id 'cd6fe08f-e9cc-4c4c-a080-61418a763276' was ended on site default. The session was ended because one of the component didn't respond in a ti
mely manner
[00:46:26] [Asr] was asked to stop listening on site default
[00:46:26] [Hotword] was asked to toggle itself 'on' on site default
[00:46:32] [AudioServer] was asked to play a wav of 91.2 kB with id '0049a91e-8449-4398-9752-07c0e1858234' on site default
[00:46:34] [AudioServer] finished playing wav with id '0049a91e-8449-4398-9752-07c0e1858234'

So notice that last bit? That is where I sent audio to be output directly to the snips server

I have a script that calls something like this to output sound

mosquitto_pub -h <MY_MQTT_IP> -t ‘hermes/audioServer/default/playBytes/0049a91e-8449-4398-9752-07c0e1858234’ -f “/tmp/sounds/temp.wav”

So you can test this to see if snips will output sound

Check here for some more details


ERROR:snips_audio_server_hermes