2025.5.0 Esphome wake word merge for s3-box-3 with sensors#163
2025.5.0 Esphome wake word merge for s3-box-3 with sensors#163chrisdunnname wants to merge 7 commits intoBigBobbas:mainfrom
Conversation
Refactor code to align even more closely with esphome wake word voice assistants core Move display changes into display script Improved behaviour for continued conversation Added progress bar for timer countdown Remove unnecessary wake word model code. Updated Voice Test Button Added hidden factory reset (10 second long press). All changes from BigBobbas#160: AT581x Radar Configuration QR Code for WIFI Enhancements for ESPHome 2025.5.0 TTS streaming resolved. BigBobbas#153 Temperature Unit Options BigBobbas#158 Code Layout improvements for flexibility Closer alignment to ESPHome S3 code Enabled all models as model can now be selected. Add climate page to status buttons
|
This is a serious refactor extending significantly on #160 to take advantage of all the updated esphome wake word voiceassistants functionality for the s3-box-3. This moves the display code into a display script so that continued conversation is much smoother in the UI and adds a progress bar for timers but keeps all the awesome UI display work done in the original from BigBobbas. The sequencing of code has changed to align more closely with the esphome scripts with one or two exceptions to avoid making it hard to merge existing configs and keeps all the ids of elements similar so any UI functions should continue to work. I couldn't test external audio yet but it should not have changed. |
Decrease to 100ms
Stop Timer on Red Button Set Media State for Player
Ensured that i2s audio speaker is finished before starting work word after announcement or media. Add logging to expedite future debugging.
Don't stop wake word on media_player state change
Display Wake Word on Device
|
i have hopefully fixed the timing with the wake sound, only tested with internal speaker currently. please create any pull requests against this file. |
|
Hi Rob, I think you found a clue that I hadn't thought of on lvgl by putting wake_sound file I had a reboot component:174]: Component i2s_audio.microphone set Error flag: unspecified |
3 participants
Refactor code to align even more closely with esphome wake word voice assistants core Move display changes into display script
Improved behaviour for continued conversation
Added progress bar for timer countdown
Remove unnecessary wake word model code.
Updated Voice Test Button
Added hidden factory reset (10 second long press).
All changes from #160 :
AT581x Radar Configuration
QR Code for WIFI
Enhancements for ESPHome 2025.5.0
TTS streaming resolved. #153
Temperature Unit Options #158
Code Layout improvements for flexibility
Closer alignment to ESPHome S3 code
Enabled all models as model can now be selected.
Add climate page to status buttons