In order to better understand and compare research techniques in building corpus-based speech synthesizers on the same data, the Blizzard Challenge has been devised.
The basic challenge is to take the released speech database, build a synthetic voice from the data and synthesize a prescribed set of test sentences. The sentences from each synthesizer will then be evaluated through listening tests.
We're pleased to announce that the Blizzard 2008 listening test is now available. We need your help in getting as many subjects as possible to participate. There are different start pages for various listener types - please take care to use the correct one:
English
======
Speech experts:
http://groups.inf.ed.ac.uk/blizzard/blizza...egister-ES.htmlVolunteers:
http://groups.inf.ed.ac.uk/blizzard/blizza...egister-ER.htmlMandarin
=======
Speech experts:
http://groups.inf.ed.ac.uk/blizzard/blizza...egister-MS.htmlVolunteers:
http://groups.inf.ed.ac.uk/blizzard/blizza...egister-MR.htmlSubjects do not need to be native speakers - we gather information about this in the questionnaire at the end of the listening test. The listening test can be completed in one session, or over multiple sessions.
If you speak both English and Mandarin, then you may participate in both listening tests if you wish.
Please publicise this as widely as possible on your mailing lists,
blogs, web pages, or whatever. Also, please ensure that as many members of your research group, other colleagues, students, family, etc
participate as possible! We need hundreds of subjects for each language. Remember to direct them to the appropriate start page.
Note that we are also running several other listener groups for paid subjects; if you would like to organise a group of paid subjects, then please contact me. We can also set up groups for other specific listener types, if we know there is a large enough pool of available subjects with certain characteristics; contact me if you want to discuss this.
Please report problems to
[email protected] 