TTS: support for Polly Neural TTS?

Add-ons Text-to-Speech
1

Hi all,

are there any plans or is maybe already somebody working on extending the Polly TTS Add-on to support the neural engine?

Neural engine: Neural voices - Amazon Polly
Neural voices: https://docs.aws.amazon.com/polly/latest/dg/ntts-voices-main.html

Currently the addon calls client.describeVoices(new DescribeVoicesRequest()).getVoices();. This already returns all voices including a parameter per voice listing supported engines:

Example:

{
      'Gender': 'Female',
      'Id': 'Vicki',
      'LanguageCode': 'de-DE',
      'LanguageName': 'German',
      'Name': 'Vicki',
      'SupportedEngines': [
        'neural',
        'standard'
      ]
},

Happy to support if nobody else is already working on it.

1 Like
2

Any traction on this topic? I have been searching for a means to implement neural voice engine myself. I haven’t found anything yet.

1 Like
3

Amazon Polly TTS now (2024) supports four voice engines. These include: Generative, Long-form, Neural, and Standard.

getVoices returns ALL voices available, but only the voices supporting the standard engine work because standard is the default when no engine is specified.

it looks like an easy addition for someone with a dev environment setup for OpenHAB

PollyTTSCloudImpl.java just does not specify an engine to use.

Currently voices are specified like this: pollytts:Joanna
I suggest adding pollytts:neural:Joanna to be able to specify which voice engine to use.

Keep support for pollytts:Joanna and default to the standard engine so we don’t break everything.

Thoughts?

1 Like