Missing language switching in Paperback even when proper language tags exist

[amirsol81]

In the attached HTM file, there is an issue with how language tags are handled in Paperback.

1. Certain characters are not displayed or read correctly.
   When the file is opened in Paperback, some characters are replaced with question marks or spoken as Unicode codes by screen readers, depending on the TTS engine in use. These same characters are displayed and spoken correctly in Bookworm and Google Chrome.
   This HTM file contains proper language tags (for example and ). In Chrome, the French sections are read with a French voice, and the English sections with an English voice, as expected.
   Paperback, however, ignores these language tags entirely and reads everything using the default English voice. As a result, pronunciation in mixed-language documents becomes unclear.
   Could you please confirm whether Paperback currently supports language tag recognition (lang= attribute) in HTML or EPUB files? If not, I would like to suggest adding this capability so that readers can automatically switch voices when encountering tagged multilingual text.
   This would make Paperback significantly more useful for language learners and anyone reading documents that include quotations or passages in different languages.

Guide.zip

[trypsynth]

Hi Amir, thanks for this issue! I'll definitely investigate the first bug. As far as your second question goes, I'm actually not sure if supporting the lang attribute in this app would do anything. I sure would like to, and if bookworm pulls it off without using a webview it's obviously possible and I just need to do some research, but to be honest I sorta just disregarded that idea when I had it come up in my brain because I have no idea how screen readers handle automatic language switching in text fields as an unfortunately monolingual person myself.

[amirsol81]

@trypsynth Thanks a lot for the reply and for checking the first issue!
About the language tags, that makes complete sense. In most screen readers, automatic voice switching only happens if the text control passes the correct lang attribute to the accessibility API, and if the synthesizer for that language is available. I understand this has not been on your list to investigate so far, and that is perfectly fine. It is just something interesting to consider whenever you have time.

[trypsynth]

Ooh okay, so yeah it's just an API I can call. Cool. I'll definitely investigate it, I won't make any promises of getting that into 0.6 but definitely pre-1.0 :)

[amirsol81]

@trypsynth Just a quick clarification: Bookworm doesn't support multilingual passages either. It's just in Google Chrome and other browsers. Bookworm handles the first issue with characters properly, though.

[trypsynth]

Ah, okay I understand. Thanks for the clarification!

[amirsol81]

Ooh okay, so yeah it's just an API I can call. Cool. I'll definitely investigate it, I won't make any promises of getting that into 0.6 but definitely pre-1.0 :)

Thanks! Well, if it materializes, Paperback will be the first accessible Windows reading app to do it as QRead and Bookworm don't support it.

[amirsol81]

@trypsynth I'm happy to report that the latest development build has fixed the character display issues. So this issue should be half-closed - haha!
The only remaining one is language identification issues even when proper language tags exist in HTML documents.

[amirsol81]

@trypsynth I edited both the issue title and issue description to remove the character-related problem and focus on the remaining language tags/ identification problem.