haliphax / tts Goto Github PK

Should be able to set this up such that the reward redemption can use a static site, able to be centrally hosted on GitHub Pages. The advanced usage (broadcast endpoint) would still require hosting a server application yourself, as this is beyond the capability of Pages.

The first execution of speech synthesis does not respect the language settings. May need to push an inaudible test first before setting the language for use in the overlay.

Currently, you have to cause a TTS event then reload the overlay entirely and cause another for it to "stick".

pactl load-module module-native-protocol-unix socket=/tmp/pulseaudio.socket
Failure: No such entity

I don't have /tmp/pulseaudio.socket, on my system with pipewire it is an empty folder

As per the instructions on https://docs.pipewire.org/page_module_protocol_native.html I modified /usr/share/pipewire/pipewire.conf to setup a unix socket which appears at /run/user/1000/pipewire-0

Tried setting default-server = unix:/run/user/1000/pipewire-0 in the dockerfile and re-build etc but I still get the below error:

tts-engine-1  | INFO:root:Speak: test (variant: m3 [random])
tts-engine-1  | Connection failure: Connection refused
tts-engine-1  | pa_context_connect() failed: Connection refused

I tried doing chmod 777 /run/user/1000 -R but still doesn't work, not sure what to do next

peter@peter-pc:~/.local/src/tts$ sudo docker-compose build
nginx uses an image, skipping
Building engine
Sending build context to Docker daemon 245.8kB
Error response from daemon: dockerfile parse error line 7: unknown instruction: SET
ERROR: Service 'engine' failed to build : Build failed

OS: LMDE 6 (faye) x86_64
Kernel: 6.1.0-13-amd64