Support for Read Aloud automatic language selection?

[readx]

Thanks for the Read aloud fix.
It works fine on Linux Mint and Windows 11, but reading in different languages (en, de, sp) requires manually changing the language output in Read aloud after starting the output.

Could a feature be added to automatically select the output language from language tags, dc:language or xml language?
Or is there another way that I just missed?

[kovidgoyal]

Use the tool to create a Read aloud audio overlay it will automatically switch voices for langauges. You can then listen to the audio overlay.

[VeryWellAged]

kovidgoyal, Thank you for ReadAloud scripting!

Please, it is possible I need Filipino English voices.

Also, if it is possible in a future release, sometime... A way to add tags for voices in an original docx, that gets used by Calibre during conversion. Here's why I ask...

With dialogue between parties, each paragraph for many pages is a different voice. Sometimes ir is two people talking, sometimes three or four. more than one-hundred-and fifty-thousands words maybe half of than is dialogue. To script it the way you have it now... well that's a bit of a nightmare. (Especially as there will be text revisions in the original doc from time to time.

[kovidgoyal]

If you want to add voices, you need to request them from the piper project. As for adding voice markup to a word document, I dont know of anyway to do that, sorry. What you can do is use a different named style in word for each character. Then when you convert this will become a unique class in calibre. You can then use the transform html tool to add the voice attribute to all tags with a specific class.

[VeryWellAged]

Originally Posted by kovidgoyal

If you want to add voices, you need to request them from the piper project. As for adding voice markup to a word document, I dont know of anyway to do that, sorry. What you can do is use a different named style in word for each character. Then when you convert this will become a unique class in calibre. You can then use the transform html tool to add the voice attribute to all tags with a specific class.

Thank you for the incredibly useful answer! As to styles... I thought I can do that. It wold have been cleaner that tags. But, no, it doesn't work ... Styles affect the entire paragraph and in a sentence like the following there are two speakers, a male and a female:
Pulling back, she smiles and says, OK, show me, Ira.

So no, styles do not work. It is not the first time I have run into the same issue with styles, as the limiation that they must impact an entire paragraph is a real problem. I wold have to create a book that could not be read but only listened to!

I will have to learn how to use the transform tool.

The Piper Project ... that's rhasspy / piper on GitHub, right? I've got no experience working with GitHub stuff, but will see what I can get done. I note that Microsoft's ReadAloud stuff has Filipio/English voices.

So and best wishes.

[kovidgoyal]

You can generate character styles not just paragraph styles.

[VeryWellAged]

Originally Posted by kovidgoyal

You can generate character styles not just paragraph styles.

I just learned that! I will give that a try!