Unifying Text to Speech from Google and Microsoft (NabuCasa)

Month of "What the heck?!"
1

In HomeAssistant, after installing the system, we have the TTS (Text to Speech) function from Google. In the location: Media > Text to Speech we can test the text-to-speech mechanism from Google. Unfortunately, we cannot choose the language and type of voice at this point. We enter the text and listen to the sound of the speech in English in a variant of a female voice.

After connecting the HomeAssistant instance with NabuCasa.com, we can also use voices, the TTS system from Microsoft. In the location: Media > Text to Speech next to “Google” is displayed the option “Cloud”. Here we can test the sound of the voice, the TTS mechanism from Microsoft. Language versions are available, we can also choose the sound of the voice.

  • Unifying: Possibility to test different voices in different languages ​​provided by Google (same as with Microsoft voices after connecting HomeAssistant with NabuCasa.com).

  • Naming: Continued naming of TTS engines according to the engine manufacturer (Google and Microsoft). We currently have Google and Cloud (this is misleading).

So much for minor tweaks to the TTS. :slight_smile:

2

Isn’t the TTS from Nabu Casa different from the Microsoft one?

3

NabuCasa.com offers TTS from Microsoft. The Polish language in Microsoft Azure is Agnieszka and Zofia’s female voices and Marek’s male voices. NabuCasa provides the voice of Agnieszka and Marek (one female and one male voice) for my language. Zofia’s voice is still available in Microsoft Azure.

You can test the available Microsoft voices at the link below: Text to Speech – Realistic AI Voice Generator | Microsoft Azure

1 Like
4

Who is missing the language and voice variant for TTS from Google in Media > Text to Speach > Google? Same as for Media > Tekst to Speach > Cloud?

Maybe there will be some votes in favor? :slight_smile:

5

I use a different AI voiceover generator online for my accessibility needs…can I unify that with Alexa. I had some ideas but it had several bug issues. As Alexa doesn’t support TTS in different natural voices I thought of using this instead :woman_shrugging:t2: