Mentioning a new Group Note from the I18N WG (Text-to-Speech Rendering of Electronic Documents Containing Ruby: User Requirements)

[murata2makoto]

Section

https://www.w3.org/TR/epub-34/#sec-tts

Describe the problem

The Internationalization Working Group has published the first draft of a Group Note titled Text-to-Speech Rendering of Electronic Documents Containing Ruby: User Requirements.

I would like to suggest another note in 9.3.2.4 Text-to-speech rendering for referencing this Note as an informative reference.

Describe the fix or new feature you propose

No response

[iherman]

Thanks @murata2makoto.

Another possibility is to add a reference into the EPUB 3 Text-to-Speech Support document, which is referred to (via a note) from the epub3 spec for more TTS related information in general.

I have no strong preference, I would let @mattgarrish (who edited that document as well) decide.

[iherman]

Note that, as a special bonus 😀, modifying the intermediate reference to the EPUB 3 TTS Support document would also provide a reference chain for the RS document. Otherwise the same reference should also be added there (into §9.3.2).

[murata2makoto]

Another possibility is to add a reference into the EPUB 3 Text-to-Speech Support document, which is referred to (via a note) from the epub3 spec for more TTS related information in general.

I am against this possibility, since the use of ruby for TTS and that of SSML/PLS for TTS are orthogonal. EPUB 3 Text-to-Speech Enhancements 1.0 is about the latter, while my note is about the former.

[iherman]

The explanation note (with the reference) in the EPUB spec says:

See EPUB 3 Text-to-Speech Support [epub-tts-10] for more information about using TTS technologies in EPUB publications.

There is no exclusivity on SSML or Lexicons in that statement. Maybe the referenced WG Note can be enhanced with a separate section on i18n issues.

[murata2makoto]

The explanation note (with the reference) in the EPUB spec says:

See EPUB 3 Text-to-Speech Support [epub-tts-10] for more information about using TTS technologies in EPUB publications.

Now, we have Aural rendering, which is about the use of TTS technologies in EPUB publications It would be nice if the explanation note mentions SSML, PLS, and CSS Speach rather than TTS technologies in general.

[mattgarrish]

Based on the discussions we just had about the TTS section in the reading system specification, I would hold off on adding this sort of informative guidance for the future TTS implementation note. I had originally proposed some general recommendations for read aloud like ensuring that image alt gets read, but the group prefers to handle these more detailed rendering requirements in the note. This case sounds the same. The recommendations for read aloud are limited to reflecting the same sort of general playback functionality that media overlays provides.

[murata2makoto]

Hi Matt,

Thanks for the clarification — that helps.

I agree that detailed rendering behavior should not be added as new requirements, and I’m not suggesting that. The ruby TTS note is intentionally informative and does not define conformance or prescribe a single correct behavior.

One possible middle ground might be to reference it very lightly from EPUB Reading Systems 3.4, rather than from the EPUB publication specification. In particular, it could be mentioned as a short Note within 9.1 Text-to-Speech, simply to acknowledge that some text interpretation issues in TTS are language- and script-specific, with ruby in CJK publications being a representative example.

This would not extend the scope of the current TTS section beyond general playback functionality, nor would it replace a future TTS implementation note. It would just point implementers to an existing analysis of a well-known interoperability issue, without introducing any new requirements.

[murata2makoto]

This issue is likely to be discussed in the context of I18N horizontal review. I18N WG knows this note very well.