Howdy, maybe this is the same issue i had with fenrir.fenrir outputs its sound to a pulse instance as root and orca fires the sound via the pulse instance of the user. i wired up a small setup script that configures the root pulseaudio instance to stream its sound to the user instance.
so both can talk at same time without locking a hardware. https://github.com/chrys87/fenrir/blob/master/tools/configure_pulse.shit needs to be run once as user and once as root to configure both parts. then a restart is requred to bring all pulse intances on pair with the new config.
maybe this helps in anyway cheers chrys Zitat von Samuel Thibault <sthibault@debian.org>:
Michael A Ray, le ven. 04 mai 2018 13:18:08 +0100, a ecrit:Perhaps a way to try is to run speech-dispatcher in TCP socket mode in a docker container.Well, I don't think a container is needed, you can just run speech-dispatcher as a system service. And the problem is not there, but rather: how will it be able to access the sound card? Unless it is running as the same user as other pulseaudio sources, there will be a conflict on the grabbing of the card.But one problem I anticipate is the way to persuade Orca that SD is there and does not need to be launched by it.That is already handled by speech-dispatcher when configure appropriately. Samuel