Voice Preview No TTS

Hello to everyone,

I’ve had my voice preview edition a few weeks but I’m yet to get it to work and I’m at the end of any diagnostics and forum suggesstions from others.

My environment:
Home Assistant in a Docker Container on a Linux Server. Latest release.
Container is in host network mode. Network is flat. The preview is on the same network as the server and the same IPv4 subnet.
Nabu Casa subscribtion (for remote and Assist Cloud).
Assist pipeline is all Home Assistant Cloud using English/British English. I’ve deleted and re-created that several times.

Symptoms and Tests
Using the Try Voice in Assist settings works fine.
Playing an MP3 through the hardware works (seemingly ruling out bad hardware).
Followed the reset to defaults process now several times. Never any issues getting the hardware connected.
Deleted the Assist pipeline several times.
Tried lots of different voices and lanaguages.
Checked server health such as disk space etc.

The Debug screen for the Assist pipeline always appears healthy. It looks like it thinks everything is fine.
I have checked the common issue of the URL for HA being incorrect but unfortuately that doesn’t appear to be it.

I have got to the point of packet captures on the network and this does hint at an issue as follows.

I can see from the packet capture the voice preview hardware is requesting the TTS audio file from HA. It opens a TCP connection to the correct LAN IPv4 address of the HA server on the port we all know and love 8123. It then sends an HTTP get request to /api/tts_proxy/xxxx.flac. From what I have read this is correct behaviour.

However it does not get a reply. If I quickly repeat that command on the server using wget I get an error of no data received.

This leads me to think that the tts_proxy component on my install is the problem. I can find nothing on where to go with that. Can anyone help?

Thank you all for any support you can give.