Customizable wakeword & timer finished sounds selection in HA for the LVA? , rather than a config.json file edit?

[so8488943-star]

Hi Imonlinux,
Any chance we can have this as a drop down selection or point and click to a custom wav/flac/mp3 file (in the HA /media? folder) via the Configuration section for the LVA in HA, rather than an edit of the config.json file?

`

"name": "My Linux Voice Assistant", --   // Optional: Wake word triggered sound (relative to repo root) "wakeup_sound": "sounds/wake_word_triggered.flac",   // Optional: Timer finished sound (relative to repo root) "timer_finished_sound": "sounds/timer_finished.flac", `

[imonlinux]

Hey @so8488943-star, let me take a look.

[imonlinux]

@so8488943-star, I have made the following changes to the thinking_sound branch.

Add event sounds, thinking sound, stable device identity, and MQTT sound selection
Event sounds & thinking sound

Add thinking_sound config option with looping support (thinking_sound_loop)
Add event_sounds_enabled master toggle gating wakeup and thinking sounds
Timer alarm intentionally not gated — functional alert
Thinking sound loops via done_callback pattern, naturally interrupted by TTS

Stable device identity — Enhancement 1

Persist MAC address to preferences.json on first boot
Subsequent boots use persisted MAC regardless of hardware changes
Fixes device re-registration when OS changes system MAC
Fix: mqtt_controller.py was ignoring passed MAC and calling uuid.getnode() independently
Fix: zeroconf.py now accepts stable MAC as parameter
Identity reset: remove mac_address from preferences.json

MQTT sound selection — Enhancement 2

Three new select entities: Sound Wakeup, Sound Thinking, Sound Timer
New switch entity: Sound Thinking Loop
Sound files scanned from sounds/wakeup/, sounds/thinking/, sounds/timer/ at startup
Subdirectories auto-created if missing
Wakeup and Thinking selects include "None" to disable; Timer does not
Selection persisted to preferences.json, applied at runtime without restart
Precedence: MQTT selection > config.json > config.py defaults
MQTT entity names prefixed (LED/Sound/Alarm) for visual grouping in HA

Bug fix

Disable mpv ytdl_hook to prevent 20-second timeout on HA TTS proxy URLs

Files changed: config.py, config.json.example, models.py, main.py, satellite.py, mqtt_controller.py, zeroconf.py, mpv_player.py
Repo structure change: Sound files moved into category subdirectories under sounds/

If you pull these changes, the sounds/ folder will include the needed sub-directories. If you have custom files in the root of your sounds/ folder, you will need to move them to the correct sub-directories to be able to list and select them in HA.

[so8488943-star]

Thanks for all the exceptional work again Imonlinux. I will play next week. Will let you know

[imonlinux]

I have been doing a lot of testing with this code branch. I have proceeded with a merge into main.

Feel free to open a issue if you find anything that doesn't work as expected. Thanks again for the suggestion!

[so8488943-star]

Hi Imonlinux,

That's fantastic!. I'll pull the latest bits down this upcoming week and have a play.

IMHO, I think this LVA fork far surpasses the OHF work, I do hope OHF merge this in and include your excellently written install guides too.

I didn't find an 'Upgrade Guide' but I guess its sufficient to do a pull of the latest bits and exec the script/setup again?
(I've read your caveats around the 'sounds' folder and I don't have a problem with that).

Just making sure , as I don't want to blow away my fantastic LVA (even though I have made an image of it jic).

Maybe an item 10. 'Upgrades' would be useful in your excellent Installation Guide. You could put caveats there per rel etc if any steps need to be carried out, beyond the normal steps.

Thanks again!

[imonlinux]

@so8488943-star, great idea!

I have updated the tutorial to include both re-installation of the reSpeaker 2Mic drivers after a kernel update and a safe LVA upgrade procedure.

I also re-ordered the Section 5 entries to put hardware installs ahead of software.

Thanks for the feedback on the project, (and the recommendations for enhancements). I had attempted to include some PR's for the OHF project, but they were not well received. Not sure if it is because I'm using GenAI to do the coding work (vibe coding), but I just don't have the patients to wait for the features that I want for my devices.

I even offered up my documentation and all I heard was crickets. To each their own...

[so8488943-star]

Thanks Imonlinux, I also put a polite request in on the OHF LVA project to merge your work . I think you've left them behind, somewhat in your dust trail. Your work is very much appreciated. I'll probably get some time to integrated your recent enhancements in the next couple of days, will keep you posted.

[so8488943-star]

Imonlinux. Step 10 do u mean 550mb? Not 550gb?

Just had a quick read.

[imonlinux]

Opps. Good catch.