Can only use "okay nabu" one time, and some error about a migrate to use ....,

[JyskMakerDK]

Hey.
Grate setup but with some compiling errors with the news update, and i can only say Okay nabo ones, and then its loos the "microfone" and there is no sound comming from it and no text if that is turned on.

I dont know how to fix it, not that into yaml. Sorry

Compiling .pioenvs/esp32-s3box-3/src/main.cpp.o
In file included from src/esphome/components/adc/adc_sensor.h:9,
                 from src/esphome.h:3,
                 from src/main.cpp:3:
/data/cache/platformio/packages/framework-espidf/components/esp_adc/deprecated/include/esp_adc_cal.h:17:2: warning: **#warning "legacy adc calibration driver is deprecated, please migrate to use esp_adc/adc_cali.h and esp_adc/adc_cali_scheme.h" [-Wcpp]
   17 | #warning "legacy adc calibration driver is deprecated, please migrate to use esp_adc/adc_cali.h and esp_adc/adc_cali_scheme.h"**
      |  ^~~~~~~
In file included from src/esphome/components/adc/adc_sensor.h:10:
/data/cache/platformio/packages/framework-espidf/components/driver/deprecated/driver/adc.h:19:2: warning: #warning "legacy adc driver is deprecated, please migrate to use esp_adc/adc_oneshot.h and esp_adc/adc_continuous.h for oneshot mode and continuous mode drivers respectively" [-Wcpp]
   19 | #warning "legacy adc driver is deprecated, please migrate to use esp_adc/adc_oneshot.h and esp_adc/adc_continuous.h for oneshot mode and continuous mode drivers respectively"
      |  ^~~~~~~

[01:48:38][D][micro_wake_word:325]: Detected 'Okay Nabu' with sliding average probability is 0.99 and max probability is 1.00
[01:48:38][D][micro_wake_word:370]: Stopping wake word detection
[01:48:38][D][micro_wake_word:378]: State changed from DETECTING_WAKE_WORD to STOPPING
[01:48:38][D][micro_wake_word:273]: Inference task is stopping, deallocating buffers
[01:48:38][D][micro_wake_word:278]: Inference task is finished, freeing task resources
[01:48:38][D][micro_wake_word:378]: State changed from STOPPING to STOPPED
[01:48:38][D][speaker_media_player:406]: State changed to ANNOUNCING
[01:48:38][D][speaker_media_player.pipeline:114]: Reading FLAC file type
[01:48:38][D][ring_buffer:034][ann_read]: Created ring buffer with size 1000000
[01:48:38][D][speaker_media_player.pipeline:124]: Decoded audio has 1 channels, 16000 Hz sample rate, and 16 bits per sample
[01:48:38][D][ring_buffer:034]: Created ring buffer with size 9600
[01:48:38][D][resampler_speaker:064]: Starting resampler task
[01:48:38][D][resampler_speaker:085]: Started resampler task
[01:48:38][D][ring_buffer:034][sample]: Created ring buffer with size 3200
[01:48:38][D][speaker_mixer:316]: Starting speaker mixer
[01:48:38][D][speaker_mixer:324]: Started speaker mixer
[01:48:38][D][i2s_audio.speaker:136]: Starting
[01:48:38][D][i2s_audio.speaker:141]: Started
[01:48:38][D][ring_buffer:034][speaker_task]: Created ring buffer with size 48000
[01:48:39][D][speaker_media_player:406]: State changed to IDLE
[01:48:39][D][speaker_mixer:329]: Stopping speaker mixer
[01:48:39][D][resampler_speaker:090]: Stopping resampler task
[01:48:39][D][resampler_speaker:095]: Stopped resampler task
[01:48:39][D][voice_assistant:477]: State changed from IDLE to START_MICROPHONE
[01:48:39][D][voice_assistant:484]: Desired state set to START_PIPELINE
[01:48:39][D][voice_assistant:207]: Starting Microphone
[01:48:39][D][ring_buffer:034]: Created ring buffer with size 16384
[01:48:39][D][voice_assistant:477]: State changed from START_MICROPHONE to STARTING_MICROPHONE
[01:48:39][E][i2s_audio.microphone:489]: Driver failed to start; retrying in 1 second
[01:48:39][E][component:286]: i2s_audio.microphone set Error flag: unspecified
[01:48:39][D][i2s_audio.speaker:148]: Stopping
[01:48:39][D][i2s_audio.speaker:154]: Stopped
[01:48:40][E][component:313]: i2s_audio.microphone cleared Error flag
[01:48:40][D][voice_assistant:477]: State changed from STARTING_MICROPHONE to START_PIPELINE
[01:48:40][D][voice_assistant:228]: Requesting start
[01:48:40][D][voice_assistant:477]: State changed from START_PIPELINE to STARTING_PIPELINE
[01:48:40][D][voice_assistant:499]: Client started, streaming microphone
[01:48:40][D][voice_assistant:477]: State changed from STARTING_PIPELINE to STREAMING_MICROPHONE
[01:48:40][D][voice_assistant:484]: Desired state set to STREAMING_MICROPHONE
[01:48:51][D][sensor:104]: 'battery_voltage': Sending state 4.26284 V with 1 decimals of accuracy
[01:48:51][D][sensor:104]: 'Battery level': Sending state 100.00000 % with 0 decimals of accuracy
[01:48:57][D][sensor:104]: 'up_sens': Sending state 162.28700 s with 0 decimals of accuracy
[01:48:58][D][light:052]: 'LCD Backlight' Setting:
[01:48:58][D][light:069]: Brightness: 30%
[01:48:58][D][light:103]: Transition length: 0.1s
[01:49:04][D][sensor:104]: 'WiFi db': Sending state -38.00000 dBm with 0 decimals of accuracy
[01:49:04][D][sensor:104]: 'WiFi Signal': Sending state 100.00000 % with 0 decimals of accuracy

[BigBobbas]

Hi can you paste the actual error please.
Thanks

[JyskMakerDK]

Hey
You can see it in the list about, i dont have any more, its just like stop lisenining.

[mbeilste]

I'm seeing the same thing. I have the ESP flashed, have installed it in HA, and the wake word will only work once and the ESP fails to process any audio after the wake word. The only way to get it to detect the wake word again is to return to the home screen, mute the mic, then unmute it via the touch screen icon. My log is very similar and also displays the microphone driver failed to start errors as well. Not sure what to do next. My ESP32-Box-3's work fine when flashed with the standard HA software.

[13:59:47][D][voice_assistant:456]: State changed from STREAMING_MICROPHONE to STOP_MICROPHONE
[13:59:47][D][voice_assistant:463]: Desired state set to IDLE
[13:59:47][D][voice_assistant:456]: State changed from STOP_MICROPHONE to STOPPING_MICROPHONE
[13:59:47][D][voice_assistant:456]: State changed from STOPPING_MICROPHONE to IDLE
[13:59:47][D][binary_sensor:036]: 'Presence detect': Sending state OFF
[13:59:48][D][light:036]: 'LCD Backlight' Setting:
[13:59:48][D][light:051]: Brightness: 100%
[13:59:48][D][light:085]: Transition length: 0.1s
[13:59:48][D][binary_sensor:036]: 'status_bar_4': Sending state ON
[13:59:48][D][binary_sensor:036]: 'status_bar_4': Sending state OFF
[13:59:48][D][switch:020]: 'Mute' Toggling OFF.
[13:59:48][D][switch:055]: 'Mute': Sending state OFF
[13:59:48][D][micro_wake_word:360]: Starting wake word detection
[13:59:48][D][micro_wake_word:378]: State changed from STOPPED to STARTING
[13:59:48][D][ring_buffer:034][mww]: Created ring buffer with size 3840
[13:59:48][D][micro_wake_word:261]: Inference task has started, attempting to allocate memory for buffers
[13:59:48][D][micro_wake_word:266]: Inference task is running
[13:59:48][D][micro_wake_word:378]: State changed from STARTING to DETECTING_WAKE_WORD
[13:59:49][D][binary_sensor:036]: 'Presence detect': Sending state ON
[13:59:50][D][binary_sensor:036]: 'Presence detect': Sending state OFF
[13:59:53][D][sensor:098]: 'WiFi db': Sending state -35.00000 dBm with 0 decimals of accuracy
[13:59:53][D][sensor:098]: 'WiFi Signal': Sending state 100.00000 % with 0 decimals of accuracy
[13:59:53][D][micro_wake_word:325]: Detected 'Okay Nabu' with sliding average probability is 0.99 and max probability is 1.00
[13:59:53][D][micro_wake_word:370]: Stopping wake word detection
[13:59:53][D][micro_wake_word:378]: State changed from DETECTING_WAKE_WORD to STOPPING
[13:59:53][D][micro_wake_word:273]: Inference task is stopping, deallocating buffers
[13:59:53][D][micro_wake_word:278]: Inference task is finished, freeing task resources
[13:59:53][D][micro_wake_word:378]: State changed from STOPPING to STOPPED
[13:59:53][D][speaker_media_player:406]: State changed to ANNOUNCING
[13:59:53][D][ring_buffer:034][ann_read]: Created ring buffer with size 1000000
[13:59:53][D][speaker_media_player.pipeline:114]: Reading FLAC file type
[13:59:53][D][speaker_media_player.pipeline:124]: Decoded audio has 1 channels, 16000 Hz sample rate, and 16 bits per sample
[13:59:53][D][ring_buffer:034][sample]: Created ring buffer with size 3200
[13:59:53][D][ring_buffer:034]: Created ring buffer with size 9600
[13:59:53][D][resampler_speaker:064]: Starting resampler task
[13:59:53][D][resampler_speaker:085]: Started resampler task
[13:59:53][D][speaker_mixer:316]: Starting speaker mixer
[13:59:53][D][speaker_mixer:324]: Started speaker mixer
[13:59:53][D][ring_buffer:034][speaker_task]: Created ring buffer with size 48000
[13:59:53][D][i2s_audio.speaker:136]: Starting
[13:59:53][D][i2s_audio.speaker:141]: Started
[13:59:54][D][speaker_media_player:406]: State changed to IDLE
[13:59:54][D][speaker_mixer:329]: Stopping speaker mixer
[13:59:54][D][resampler_speaker:090]: Stopping resampler task
[13:59:54][D][resampler_speaker:095]: Stopped resampler task
[13:59:54][D][voice_assistant:456]: State changed from IDLE to START_MICROPHONE
[13:59:54][D][voice_assistant:463]: Desired state set to START_PIPELINE
[13:59:54][D][voice_assistant:186]: Starting Microphone
[13:59:54][D][ring_buffer:034]: Created ring buffer with size 16384
[13:59:54][D][voice_assistant:456]: State changed from START_MICROPHONE to STARTING_MICROPHONE
[13:59:54][E][i2s_audio.microphone:489]: Driver failed to start; retrying in 1 second
[13:59:54][E][component:195]: Component i2s_audio.microphone set Error flag: unspecified
[13:59:54][D][i2s_audio.speaker:148]: Stopping
[13:59:54][D][i2s_audio.speaker:154]: Stopped
[13:59:55][E][component:209]: Component i2s_audio.microphone cleared Error flag
[13:59:55][D][voice_assistant:456]: State changed from STARTING_MICROPHONE to START_PIPELINE
[13:59:55][D][voice_assistant:207]: Requesting start
[13:59:55][D][voice_assistant:456]: State changed from START_PIPELINE to STARTING_PIPELINE
[13:59:55][D][voice_assistant:478]: Client started, streaming microphone
[13:59:55][D][voice_assistant:456]: State changed from STARTING_PIPELINE to STREAMING_MICROPHONE
[13:59:55][D][voice_assistant:463]: Desired state set to STREAMING_MICROPHONE

[mbeilste]

I did some more digging and the single wake issue was in HA with the voice assistant. It took several tries to get it setup and it is now working correctly. I believe this issue can be closed and is related to the HA side, not the firmware in the ESP32. I do know that the mic driver error causes some strange/awkward timing issues when talking with the device. If you don't pause to let the error clear, the device will not catch what you say. The stock HA firmware does not throw this error and you can rattle off commands much faster and more reliably. Just load up the wireless log and give commands to see the errors in real time. If you talk too quick, it will miss your commands since the error has not yet cleared.

[galsom]

@mbeilste can you please give us some more insights into what you changed in HA?

I played around with local wake word detection in HA versus on device, but fail to get this working.

[mbeilste]

I'm not 100% sure which step I did that solved it since its been a few weeks but I remember reflashing the ESP32 and playing around with the assistant setup under the device area (not the drop downs, but the wizard that runs to setup the ESP's audio assistant). This would be found under HA under the Settings->Integrations->ESP Home area. There are three dots to click and then select "Setup Audio Assistant". I also did a lot of playing around with different assistants, but I think if you can start a conversation under the assistant area of HA and it runs OK your assistant should be fine.

[galsom]

Thank you for the feedback. I'll continue to play around as well.

[tepebox]

Try updating to the lastest version of ESPBuilder... Did the trick for me with a similar issue...