No sound played, but it is generated - TTS, Logitech Media Server, max2play

Configuration
1

Hi,

HASSIO running on Ubuntu 16.04.6 host
MaryTTS in docker - synesthesiam/marytts:5.2
Max2Play Logitech Media Server (LMS) running on a raspberrypi

I have a strange issue in that sending a TTS from HASSIO to LMS isn’t being played.
I’m not running any SSL as I’ve read that was stopping LMS get the media file from HTTPS urls.
The URL shows up in LMS and it looks to begin playing but no sound is heard. If I download the file using the URL passed to LMS I can play the file, I can upload it to LMS media library and have LMS play it.

I’d apreciate any pointers here as I’m at a totall loss where to go next.

2

Google TTS worked, really want a local or maybe static audio files solution so would like to get marytts working.

3

So, I have an answer, that might be close to the answer.

LMS (Logitechmediaserver) is not reading what MaryTTS is sending, either Mary is not setting the mime type correctly or LMS is not reading it.

However, there is a way! LMS allows us to tell it what file extensions map to what file types.

In the config directory alongside the default LMS types.conf file create a file called custom-types.conf with the following line:

wav     wav,wave,wave_file audio/x-wav,audio/wav,audio/vnd.wave,audio/wave     audio

Here I am overriding the wav file line in types.conf with an identical line except I added wave_file as a wave file type.
However I got some evil feedback a couple of times where I believe a junk file was served by homeassistant or mary, or possibly homeassistant sent an html error which was read as a WAV file!

I will test things a bit more as I go on. Hope this helps.