To my knowledge this is a subject that has still to be cracked,all mods use existing vocals or have blanks for their new bases and systems.

**shuffles of with a new headache**

Cracking it is not that hard, the UTF editor will export and import wav files into the original speech files with no problems. The trouble is the scope of such a project. There are dozens of speech files broken into hundreds of wav files each. To mod it, one would need to work out the script to every sound file and then create an alternative script that would have to be rigidly adhered to. It is an aweful lot of work and most modders have more important things to do with their time then work on something that big. We did look into doing it for Freeworlds, but again, the time restraint was a massive thing that is holding us back.

the entire GCS system remains a mystery really,i would focus on pulling that apart first - and i wouldn't want to try modding speech with UTF edit it's far too primitive to import dozens of wavs,with no preview to tell you where you're importing to,ugh

if you want a mass extraction of wavs you need the UTF to XML project, unfortunately they've never followed up and made an XML to UTF converter which is sad since it could really help & have real potential for making user designed MMORPG type objects

Is it impossible? No.
Is it extremly time-consuming? Yes, definitly.

I've tried to replace some of the parts of the UTF-files. And my experiment worked. I exported the file for Planet Manhattan, reversed it and imported it again. It worked perfectly. Of course the new file has to have the same format as the exported file, otherwise it won't work.

So the only problem remaining is to find good and willing speekers and patient audio editors, who get every sound file into the right format, import it, give it the right hash name and make the right entry in the sound file's INI.

I think its hostile universe that employs a couple of home made vocals on base names and the Welcome to speeches. So it can me be done.