Use of SSML in EPUB

[murata2makoto]

In the introduction of "Pronunciation Gap Analysis and Use Cases"

There are technical methods to allow authors to inline SSML within HTML (using namespaces), but such an approach has not been adopted, and anecdotal comments from browser and assistive technology vendors have suggested this is not a viable approach.

SSML is used in Japanese EPUB publications for textbooks. The biggest textbook publisher in Japan, Tokyo Shoseki, extensively use SSML. It is hard to imagine that the use of SSML for K12 digital textbook will decrease. Thus, the above sentence should be weakened.

However, I would certainly welcome generic mechanisms in HTML.

[mhakkinen]

Thank you for the comment. The use of SSML in EPUB is mentioned in the document, but I agree that the statement you highlight should be clarified.

Based on information from the DAISY Consortium, the EPUB usage of name-spaced SSML attributes is limited to the phoneme element, and there is no reading systems support [1].

It would be helpful to the TF to better understand the usage in Japan. My understanding is that the SSML attribute is only consumed in the EPUB production process (when audio files are generated), and that it is not supported in any reading systems. If you know, or can point us to someone who does know, whether any reading systems in Japan directly consume the in-line SSML attributes at reading time. Also, it would be interesting to know whether the Japanese publishers have included additional SSML attributes.

[1] http://kb.daisy.org/publishing/docs/text-to-speech/ssml.html

[mhakkinen]

@murata2makoto, I wanted to follow up and ask if you have any contact at Tokyo Shoseki who can describe specifically, and perhaps provide examples, of their use of SSML. Have they extended the use of the SSML attribute approach of EPUB beyond phoneme? Any information you have would be welcome.

Thanks!