TTS and MPD audio issues - again

Hi community,

When using tts (google + tts.speak or tts.picotts_say) with local MPD, first part of audio appear trimmed. So If I write “one two trhee four five” then I hear “our five”.

I also can see the common and old error message: “Error on load tts: Wrong tts file format!” but not sure if related.

I’m running this:

  • Hardware: raspberry pi 5
  • OS: dietpi
  • Sound card: hdmi
  • HAAS: 2024.9.1 custom install
  • MPD: 0.23.12

Any idea to debug/solve this?

Adding some at the beginning to the tts text won’t be useful, for obvious reasons.

thanks in advance !