Quote:
Originally Posted by DaleDe
One of the problems with getting too high a quality in the TTS is that the publishers and writers guild will demand that it be disabled, as they claim it cuts into sales of the audio versions of the book. It is a fine line but Amazon already crossed it. It do think Jinke should offer the user the choice of downloading voices but not provide too good a one with the device.
Dale
|
Dale I would suggest the following:
1) Jinke is not selling books.
2) Without SSML the voices while much improved on the present; they are not performances just much much better mechanical voices.
3) With SSML the work involved is not simple nor can it ever be fully automated. A well written SSML with the quality of voice and voice engine referred to, can create performed text.
4) Authors should be paid for this rendition of their work, but that has to be negotiated between them and their publisher.
Audiobooks will always have a role and human performance is no small thing no matter how good technology gets -
however - there is simply no way that all the books that could be performed would be performed -- obviously not every EPUB publisher will add this -- but what I intend to put together are EPUBS which include a TEI markup, SSML with IPA transliteration regardless of what the present generation of readers can use (plus if things go right MP3 as well).
Jinke have a slight edge at the moment, by being so well designed for purpose and not being linked to bookselling -- a good TTS approach would give it prominence and set the standard in real features rather than bells and whistles.