satellite gets stuck after only 1 wake word detection #154
Comments
|
Hi all, After compiling required libs for activating VAD; I noticed that my voice is well detected Here is the command I'm using with VAD argument: And here are the logs whowing a successful wakeword recognition: And here; same logs showing my attempts to trigger wakeword again with no luck: (we can see the audio stream is detected when I speak; so it seems it is detected by the satellite, but the wakeword never gets triggered) If I want to detect a wakeword again; I still need to kill and restart the satellite process. Then it works only once and I have to restart the process again. Do anyone has some idea to share on this so that I can get it fixed maybe ? :) NB: I'm also wondering what is "restart_on_end" that shows up in the logs ? Is it supposed to be "True" to somehow "allow" the wakeword going in "detecting" state again ? Thank you for your feedback :) |
|
I have the same issue, I installed everything yesterday. |
|
I've tried previous versions, but the result was the same. I've tried: |
|
Hi all, So the device is using the same Home Assistant setup than the one I used with wyoming-satellite: wyoming openwakeword (still with my custom wakeword) Only difference is that I replaced wyoming-satellite by an SP32-S3-BOX-3 And with this setup; I'm able to trigger the wakeword several times without having to restart anything. NB: In case it can help: I'm using docker version of Home Assistant and wyoming-openwakeword Thank you for your help |
|
Additional context: I'm following https://www.youtube.com/watch?v=eTKgc0YDCwE&list=WL&index=54 with the same hardware: Additional logs: I added Only the called commands logs: And this is my service definition: [Unit]
Description=Wyoming Satellite
Wants=network-online.target
After=network-online.target
Requires=wyoming-openwakeword.service
Requires=2mic_leds.service
[Service]
Type=simple
ExecStart=/home/diego/wyoming-satellite/script/run \
--name 'Alfred Satellite' \
--uri 'tcp://0.0.0.0:10700' \
--mic-command 'arecord -D plughw:CARD=seeed2micvoicec,DEV=0 -r 16000 -c 1 -f S16_LE -t raw' \
--snd-command 'aplay -D plughw:CARD=seeed2micvoicec,DEV=0 -r 22050 -c 1 -f S16_LE -t raw' \
--mic-auto-gain 0 \
--mic-noise-suppression 2 \
--mic-volume-multiplier 1.0 \
--wake-uri 'tcp://127.0.0.1:10400' \
--wake-word-name 'hey_jarvis' \
--event-uri 'tcp://127.0.0.1:10500' \
--awake-wav sounds/awake.wav \
--done-wav sounds/done.wav \
--tts-stop-command 'true' \
--detection-command examples/commands/detection.sh \
--startup-command examples/commands/startup.sh \
--streaming-start-command examples/commands/streaming_start.sh \
--streaming-stop-command examples/commands/streaming_stop.sh \
--synthesize-command examples/commands/synthesize.sh \
--stt-start-command examples/commands/stt_start.sh \
--stt-stop-command examples/commands/stt_stop.sh \
--tts-start-command examples/commands/tts_start.sh \
--tts-stop-command examples/commands/tts_stop.sh \
--debug
WorkingDirectory=/home/diego/wyoming-satellite
Restart=always
Hopefully @synesthesiam can take a look at this 🙏 |
|
I have the same setup, and I get into the same issue.
The LED service doesn't even work. I was never able to turn the LED on |
|
Hi everyone |
|
Similar issue here, in debug mode it only responds to wake word one time and gets stuck. The satellite is configured in Home Assistant. I'm using most of the defaults from the tutorial since I have the same hardware |
|
Same problem here, trying to fix for a couple of days now with no success. Similarly to some people on this thread, I am using this page from the docs since that is the equipment I have (raspberry pi 0 2w and respeaker 2-mic pihat). In the logs bellow you can see the wake word triggering the first time I use it, but then it stops working: My wyoming-satellite.service file: |
|
Hi @synesthesiam, Can we kindly ask you for some support on this please? 🙏 |
|
Hello, Steps I took to fix it:
Not sure if all of the above is needed, but the core of the issue for me was not having a Speech to text agent configured, which made the service stuck in listening mode. #87 I suppose this could be the same issue as I had. |
Thank you for sharing this. I will research about it and possibly try your solution today or tomorrow and report back. The logs you refer too are just the one in settings >> system >> logs? I don't seem to have any error regarding to speech-to-text in mine. DO you have any idea why these integrations need to be installed on the HA side of things? I thought that the satellite should take care of these things on their end, which is why I was trying to get it ready to go before working on the HA side of things. I am trying to figure out if we had indeed the same error |
|
First, I enabled logging on the Wyoming integration and then recreated the bug. Afterward, I went through the 'home-assistant.log' file, but it seems to be the same thing as you suggested. (Im fairly new to HA) To be honest, I'm not entirely sure how it is intended to work in Home Assistant. My guess is you only need to have any speech-to-text service set up in the voice assistant settings. |
|
It was not as straight forward for me, but it finally worked, thanks to @TomasKliner ! Unfortunately, I could not install the add-ons since there is no possibility for me to get a dedicated machine for HA, and add-ons are not supported in the docker container of HA installation. I understand all the arguments of why it is like this and although I agree with most of them, this problem proves that the argument of "but if you install on docker you are good enough to have any add-on on docker as well", which I read around github and reddit posts, is just wrong, but I digress. |
I've had the same problem and interestingly enough, as @richardoswald implies, when I removed the Edit: Cancel that, it sometimes works and sometimes doesn't without a clear pattern to me. Edit 2: After some more searching, I've got it to work via #217. I had to copy the custom wake word model to the default model location: Having to copy a file into a Python package installed in a virtual environment is obviously an ugly workaround, though. From my testing, I think there's also a second bug, that makes the Here's a docker compose file that works on my Raspberry Pi: |
|
@jerome83136 did you find a solution? I'm stuck here as well: I have no Idea how to debug it and understand where the pipeline stop. after restarting the service I can trigger the wake word again, sometimes rarely it (kind of) auto-reset. @synesthesiam :-) edit: added complete DEBUG journals |
|
After several attempt I went a bit forward: my config for wyoming satellite is But on Home Assistant side I still don't see "Assist in progress" binary sensor activating.. Someone can help me to understand? EDIT: Solved, the TTS and STT integrations were not configured. |
|
For reference, when I test this directly on the device, it doesn't work and it gets stuck (see my previous comments). However, when I use it through Home Assistant, it magically works as expected 🤷, which is how I plan to use it. |
Hello,
I built wyoming-satellite and I run it this way:
/opt/wyoming-satellite# script/run --name "my satellite" --uri tcp://0.0.0.0:10700 --mic-command "arecord -D plughw:CARD=seeed8micvoicec,DEV=0 -r 16000 -c 1 -f S16_LE -t raw" --snd-command "aplay -D plughw:CARD=seeed8micvoicec,DEV=1 -r 22050 -c 1 -f S16_LE -t raw" --awake-wav sounds/awake.wav --detection-command examples/commands/detection.sh --wake-refractory-seconds 0 --debugI'm using a custom wakeword (Google colab) and it works fine; but it only works "once"
I mean:
And that's it. I am unable to get the wakeword activating again without killing and restarting the satellite process.
After restarting; it works (for 1 time) etc...
NB: I had the same result when using standard wakewords (hey_rhasspy, hey_jarvis, etc...)
Here are DEBUG logs of my satellite:
What did I do wrong please ?
Hardware:
Thank you for your help :)
Best regards
The text was updated successfully, but these errors were encountered: