Pulseaudio and best Speech-dispatcher setup

[JanuszChmiel]

Several days Ago, MR Thibault, one of The core developers of Speech-dispatcher have introduced version 10 RC of this speech server. He has wanted to improve speech quality for Linux users, so he has changed Buffer for Pulseaudio streaming inside Speech-dispatcher source code. He has set The buffer value to 10 MS. As A result, this short buffer is causing severe crash of PUlseaudio process, which run from Termux. Crash have been occurred, when I have used Unix socket, no TCP/IP. Please, which buffer value will be The best, because 10 MS is too short buffer when Pulseaudio uses SLess module to communicate with other sound sub systems for Android.

I have also used SHMEM library developed by aauthor of Xserver XSDL. Do you think, that it would be possible to use some pulseaudio setting from Termux, or should I recompile SPeech-dispatcher and modify its internal setting from 10 MS to 20 or higher value? Older Speech-dispatcher version 1.9.2 or similar value did not caused Pulseaudio sound server auto termination. Crash will not close whol Termux session and running processes, Pulseaudio sound server terminates. Thank you very much for yours advice.