search

nirnachmani / ESP32-S3-Box-3-Firmware

ESP32 S3 Box 3 Firmware which combines continuous conversation, dock sensors and media player
6 stars 4 forks source link

voice like Mikey Mouse and fast #5

Open MrJamsn opened 5 months ago

MrJamsn commented 5 months ago

First of all, thank you for your awesome work!!

I installed it correctly but have two issues, one only since I've installed this firmware.

The response voice is fast and high pitch, like mikey mouse. Is there any setting to speed up the reponses?

(second issue, maybe it is familiar. after a couple of minutes, the wake word doe not get recodgnized anymore. only after a restart it is operational again)

thank you!

nirnachmani commented 5 months ago

For the distorted voice, try to change the sample_rate of announcement_audio in the media_player block. The value in the original code was 16000 but this gave distorted voice (kind of opposite of what you are describing). Changing it to 22050 solved the problem. Maybe for you 16000 will work instead.

I am not sure about the second problem. It doesn't happen with my unit. Try to look at the log as you say the wake word to see if something happens.

MrJamsn commented 5 months ago

Thank you, the sampling rate was the right place. 16000 works perfectly! Second problem, didn’t occur yesterday. Will check the logs.You know by any chance how to let it answer with a beep or tone if a entity was set? (Instead of “kitchenlight was turned on”)Thank you! ____Raffael Jambor, @. meinem iPhone gesendetAm 13.06.2024 um 11:55 schrieb nirnachmani @.>: For the distorted voice, try to change the sample_rate of announcement_audio in the media_player block. The value in the original code was 16000 but this gave distorted voice (kind of opposite of what you are describing). Changing it to 22050 solved the problem. Maybe for you 16000 will work instead. I am not sure about the second problem. It doesn't happen with my unit. Try to look at the log as you say the wake word to see if something happens.

—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: @.***>

nirnachmani commented 5 months ago

Great.

Regarding using a confirmation sound instead of a verbal answer, this kind of functionality is not part of the Box3 firmware. Instead, it is part of the assist pipeline that the boxes uses. It depends on what components you use in your assist pipeline. You may be able to implement a confirmation sound after you say the wake word (see https://github.com/esphome/feature-requests/issues/2490) but this is not what you are asking for.

MrJamsn commented 5 months ago

Thank you for the link! Will have a look :)

BR

Am 15.06.2024 um 06:29 schrieb nirnachmani @.***>:

Great.

Regarding using a confirmation sound instead of a verbal answer, this kind of functionality is not part of the Box3 firmware. Instead, it is part of the assist pipeline that the boxes uses. It depends on what components you use in your assist pipeline. You may be able to implement a confirmation sound after you say the wake word (see https://github.com/esphome/feature-requests/issues/2490 <x-msg://3/url>) but this is not what you are asking for.

— Reply to this email directly, view it on GitHub https://github.com/nirnachmani/ESP32-S3-Box-3-Firmware/issues/5#issuecomment-2169120589, or unsubscribe https://github.com/notifications/unsubscribe-auth/AJRE25FTCFP7J5CPOJWKD73ZHO7J7AVCNFSM6AAAAABJG46HXOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCNRZGEZDANJYHE. You are receiving this because you authored the thread.