pass options through to tts.say for voice

[Edz2K23]

When using tts.say and the piper tts service how can the option: voice be passed through to use a non standard voice?

This is so cusom voices loaded into /share/piper can be used. jsdon and onnx models can be used this way by specifying them:

  options:
    voice: jarvis-medium

Ive tried adding that to the configuration but either Ive got it wrong or its not passed through, as is piper can only speak in the default set voice which limits you to those in the container as the default (again unless im missing something)

- platform: lms_tts_notify
  name: testbox
  tts_service: tts.speak
  entity_id: tts.piper
  media_player: media_player.picoreplayer
  device_group: group.all_persons
  options:
    voice: jarvis-medium
  volume: 0.4

[Edz2K23]

So I think this can be done via the options passed into tts.speak

options: { voice: en_GB-glados-medium }

I have looked at init.py and suspect this can be passed in on line 556 but that doesnt seem to work (have tried with and without the quotes as I thought may be a python dictionary issue)

        # Play message
        if self._message:
            if 'speak' in self._tts_service:
                service_data = {
                    'entity_id': self._tts_engine,
                    'media_player_entity_id': self._media_player,
                    'message': self._message,
                    **'options': '{ voice: "en_GB-glados-medium" }',**

[floris-b]

You cannot set this in the configuration of tts.speak?
I am using google_cloud tts where is can set the voice and other options