Testing the satellite1.yaml config from develop branch - results

[ms1design]

Testing environment

Home Assistant: 2024.12.3
Music Assistant: 2.3.3
Branch: develop [385676b]
Flashed config: satellite1.yaml
XMOS Firmware: v1.0.1-alpha.40
Onboarding method: USB-C / ESPHome

Testing workflow

Add use_address to config/common/wifi_improv.yaml to enable to tail logs though WiFi:

wifi:
  id: wifi_id
  use_address: satellite1-XXXXX.local

1. Flash:

esptool.py -p /dev/cu.usbmodem101 erase_flash | tee erase.log && \
esphome clean config/satellite1.yaml | tee clean.log && \
esphome compile config/satellite1.yaml | tee compile.log && \
esphome upload config/satellite1.yaml | tee upload.log

2. Connected to WiFi though web.esphome.io
3. Tail logs using WiFi:

esphome logs config/satellite1.yaml | tee runtime.log

4. Powered using USB-C with 3A @ 9V

Issues

In rare situations the Error: stt-no-text-recognized - No text recognized is printed out in logs and all LED's blink red after starting listening to the voice command - this happens immediately and user is even unable to say the final action command (not the timestamps). Finished Speaker Detection is set to Default.

[15:01:37][D][light:109]:   Effect: 'Waiting for Command'
[15:01:38][D][voice_assistant:641]: Event Type: 11
[15:01:38][D][voice_assistant:804]: Starting STT by VAD
[15:01:38][D][light:036]: 'voice_assistant_leds' Setting:
[15:01:38][D][light:051]:   Brightness: 66%
[15:01:38][D][light:109]:   Effect: 'Listening For Command'
[15:01:39][D][voice_assistant:641]: Event Type: 12
[15:01:39][D][voice_assistant:808]: STT by VAD end
[15:01:39][D][voice_assistant:516]: State changed from STREAMING_MICROPHONE to STOP_MICROPHONE
[15:01:39][D][voice_assistant:522]: Desired state set to AWAITING_RESPONSE
[15:01:39][D][voice_assistant:516]: State changed from STOP_MICROPHONE to STOPPING_MICROPHONE
[15:01:39][D][light:036]: 'voice_assistant_leds' Setting:
[15:01:39][D][light:051]:   Brightness: 66%
[15:01:39][D][light:109]:   Effect: 'Thinking'
[15:01:39][D][voice_assistant:516]: State changed from STOPPING_MICROPHONE to AWAITING_RESPONSE
[15:01:39][D][voice_assistant:516]: State changed from AWAITING_RESPONSE to AWAITING_RESPONSE
[15:01:39][D][voice_assistant:641]: Event Type: 0
[15:01:39][E][voice_assistant:776]: Error: stt-no-text-recognized - No text recognized
[15:01:39][D][voice_assistant:634]: Signaling stop...
[15:01:39][D][voice_assistant:516]: State changed from AWAITING_RESPONSE to STOP_MICROPHONE
[15:01:39][D][voice_assistant:522]: Desired state set to IDLE
[15:01:39][D][voice_assistant:516]: State changed from STOP_MICROPHONE to IDLE

When Customer does not have the TTS/STT configured in the Assist Pipeline: The blue LED's are not lighting up when detecting Wake Word. Based on the logs it seems that the player is changing state to playing and we can hear the wake word detection sound though speaker, then it immediately switch to idle state. The assist_satellite is stuck on Idle state.

Where this is expected and not an issue, we can improve the user experience with at least a dedicated message printed in logs, eg: Your Voice Assist Pipeline is missing required components like STT or TTS.

14:34:52][D][sensor:094]: 'Ambient light': Sending state 415.22882 lx with 1 decimals of accuracy
[14:35:00][D][sensor:094]: 'Temperature': Sending state 28.26411 °C with 2 decimals of accuracy
[14:35:00][D][sensor:094]: 'Humidity': Sending state 46.23573 % with 2 decimals of accuracy


[14:35:13][D][micro_wake_word:357]: Detected 'Okay Nabu' with sliding average probability is 0.99 and max probability is 1.00
[14:35:13][D][media_player:080]: 'Sat1 Media Player' - Setting
[14:35:13][D][media_player:084]:   Command: STOP
[14:35:13][D][media_player:093]:  Announcement: yes
[14:35:13][D][media_player:080]: 'Sat1 Media Player' - Setting
[14:35:13][D][media_player:093]:  Announcement: yes
[14:35:13][D][nabu_media_player.pipeline:174]: Reading FLAC file type
[14:35:13][D][nabu_media_player.pipeline:186]: Decoded audio has 1 channels, 48000 Hz sample rate, and 16 bits per sample
[14:35:13][D][nabu_media_player.pipeline:211]: Converting mono channel audio to stereo channel audio

[14:35:36][D][micro_wake_word:357]: Detected 'Okay Nabu' with sliding average probability is 1.00 and max probability is 1.00
[14:35:36][D][media_player:080]: 'Sat1 Media Player' - Setting
[14:35:36][D][media_player:084]:   Command: STOP
[14:35:36][D][media_player:093]:  Announcement: yes
[14:35:36][D][media_player:080]: 'Sat1 Media Player' - Setting
[14:35:36][D][media_player:093]:  Announcement: yes
[14:35:36][D][nabu_media_player.pipeline:174]: Reading FLAC file type
[14:35:36][D][nabu_media_player.pipeline:186]: Decoded audio has 1 channels, 48000 Hz sample rate, and 16 bits per sample
[14:35:36][D][nabu_media_player.pipeline:211]: Converting mono channel audio to stereo channel audio

[14:35:48][D][ltr_als_ps:185]: Reconfiguring sensitivity: gain = 1x, time = 100 ms
[14:35:49][D][ltr_als_ps:185]: Reconfiguring sensitivity: gain = 2x, time = 100 ms
[14:35:50][D][ltr_als_ps:185]: Reconfiguring sensitivity: gain = 4x, time = 100 ms
[14:35:51][D][ltr_als_ps:185]: Reconfiguring sensitivity: gain = 8x, time = 100 ms
[14:35:52][D][ltr_als_ps:450]: Illuminance is sufficient.
[14:35:52][D][sensor:094]: 'Ambient light': Sending state 377.65100 lx with 1 decimals of accuracy

Very occasionally the Wake word model predicts 'Okay Nabu', but VAD model doesn't.

[14:12:29][D][voice_assistant:516]: State changed from IDLE to START_MICROPHONE
[14:12:29][D][voice_assistant:522]: Desired state set to START_PIPELINE
[14:12:29][D][voice_assistant:225]: Starting Microphone
[14:12:29][D][ring_buffer:034]: Created ring buffer with size 16384
[14:12:29][D][voice_assistant:516]: State changed from START_MICROPHONE to STARTING_MICROPHONE
[14:12:29][D][voice_assistant:516]: State changed from STARTING_MICROPHONE to START_PIPELINE
[14:12:29][D][voice_assistant:280]: Requesting start...
[14:12:29][D][voice_assistant:516]: State changed from START_PIPELINE to STARTING_PIPELINE
[14:12:29][D][voice_assistant:641]: Event Type: 2
[14:12:29][D][voice_assistant:733]: Assist Pipeline ended
[14:12:29][D][voice_assistant:516]: State changed from STARTING_PIPELINE to IDLE
[14:12:29][D][voice_assistant:522]: Desired state set to IDLE
[14:12:29][D][light:036]: 'voice_assistant_leds' Setting:
[14:12:29][D][voice_assistant:634]: Signaling stop...
[14:12:33][D][micro_wake_word:352]: Wake word model predicts 'Okay Nabu', but VAD model doesn't.
[14:12:33][D][micro_wake_word:352]: Wake word model predicts 'Okay Nabu', but VAD model doesn't.
[14:12:33][D][micro_wake_word:352]: Wake word model predicts 'Okay Nabu', but VAD model doesn't.
[14:12:33][D][micro_wake_word:352]: Wake word model predicts 'Okay Nabu', but VAD model doesn't.
[14:12:33][D][micro_wake_word:352]: Wake word model predicts 'Okay Nabu', but VAD model doesn't.
[14:12:33][D][micro_wake_word:352]: Wake word model predicts 'Okay Nabu', but VAD model doesn't.

Other observations

• when my Satellite1 finally switched to the correct WiFi mesh device the connectivity improved and finally I can stream music from Music Assistant and Voice Commands work more reliably
• when I'm playing music though Music Assistant on Speaker and I press Action Button it pauses it. When I press another time it starts the Voice Pipeline. What is the purpose of this? :)
• I would love to have sweet boot sound. Maybe I'm a bit nostalgic but https://www.youtube.com/watch?v=miZHa7ZC6Z0 always warms my heart :D

Dumped logs

compile.log
erase.log
runtime.log
upload.log

[ms1design]

@gnumpi when we stream music from Music Assistant and press the HW Mute button on Satellite1 it should mute the external Speaker?

[gnumpi]

@gnumpi when we stream music from Music Assistant and press the HW Mute button on Satellite1 it should mute the external Speaker?

No, the HW Mute button is muting the microphones only.