full-fledged pidgin

Gizmodo already plugged moanmyip.com but it’s worth mentioning that there probably really is a (commercial) research avenue here for convincing sexually-provocative speech synthesis. Maybe if the actual project I’m working on right now doesn’t pan out I’ll submit a research proposal for a project to build titillating interactive dialogue systems. ;) I mean, these guys didn’t even take the time to record multiple versions of each numeral for concatenation when numerals repeat.

Who knows, it could be the killer app for speech? Broader societal impact: providing pornographic content without exploiting any actual people (other than the customers).

Of course, the first corollary of rule 34 suggests that such technology already exists and is deployed somewhere on the Internet.

I’m currently looking for someone who might be interested in working with me on a speech synthesis research project. The gig would involve some time in the recording studio recording transcripts intended to elicit as-nearly-complete-as-possible a selection of English speech sounds in different contexts, different emotional conditions, and at different paces. Speaker should be comfortable reading copy into a mic (acting experience/training a huge plus), interested in learning something about how his/her vocal tract works, and willing to participate in a number of articulatory measurements (think tubes, electrodes, ultrasound wand placed submentally (beneath your jaw)) for surprisingly little compensation. :)

There would be two sessions:

1. Session 1: high-quality sound recording as described above.
2. Session 2: recording of same transcripts in a phonetics laboratory with various instrumentation installed and attached.

When we’re done you get an open-source speech synthesizer that talks in your voice. Please comment or e-mail me if you’re interested (or know someone who might be).