- From: David Woolley <david@djwhome.demon.co.uk>
- Date: Tue, 16 Aug 2005 07:31:52 +0100 (BST)
- To: w3c-wai-ig@w3.org
> > You may want to look at open source systems based on MBROLA / festival or > similar if you want a MÄori screen reader - I don't know if anyone has What the industry means by screen reader these days is a an MSAA/HTML document object model reader, like JAWS. That seems to be the standard by which web authors decide if they are minimally legally compliant. How many open source MSAA/DOM based tools are there? > done the work but as I understand it MÄori is written in a realtively > phonetic manner, so shouldn't be all that difficult to work with for a > voice developer. Most languages that have only recently been given an alphabet base orthography have fairly simple spelling rules, but that doesn't necessarily mean that text to speech is easy as spelling rules generally oversimplify. They tend to merge similar phonemes into a single one where there is no resulting ambiguity, and they don't, normally, fully account for sandhi, the changes in sounds when two phonemes are adjacent to each other (often a change from voiced to unvoiced or vv).
Received on Tuesday, 16 August 2005 19:23:37 UTC