Talia

So I started work on implementing my work-flow for importing voice overs and phoneme mapping them in Faceposer. I wasn't aware that Faceposer's SAPI DLLs didn't work with Windows 7, and it took me several hours to understand why it wasn't working. Luckily, Mr. Tom Edwards has spent some time on this problem already, and hacked together a DLL file that is supposed to work with the newer speech recognition libraries. This is amazing, and awesome.
Unfortunately, it still appears to be somewhat glitchy. I blame this on Microsoft. Oddly, MS is usually always pretty good about including backwards compatability on a product. I've rarely if ever had issues with a minor version update (5.1 to 5.4 of SAPI) causing this sort of regression, so I'm a bit confused, to be honest. But it definately does not have the same hooks, it appears. Tom's version of the DLL is causing it to output the right phone time mappings, but it's outputting Silence phonemes for me, basically reporting that it doesn't know what to map the phones to.
So I'm trying to work around it for the moment, and get on with the movement bits of the script. It's just nice to have in-situ audio so that way I know approximately how it should play out. It will be a bigger deal obviously when I get closer to the end and need to add final audio in. Right now I'm just doing the voice overs myself, so I can see how it plays out.
I'm mostly amazed at the amount of work some people outside of valve have put in to get some of these things working. It's pretty awesome, actually.