Ok. So a bit more information. I apologise if any of this is obvious.
This is the guide I followed: http://wiki.tesnexus..._Complete_Guide
I put the facefxwrapper in tools/audio
Then Data/Sound/Voice/Processing I have the FonixData.cdf - (I think I downloaded it from nexus)
From memory the above was the fiddly bit. I am afraid I can not tell you how this got set up properly. I vaguely remember that the fonixdata.cdf file did not seem to work by itself and that facefxwrapper was needed. I am sorry I am a bit sketchy here. I just can not remember how this got sorted. CK kept throwing up errors but then at some point it started to work.
The below can all be found in the guide above:
In the Voice folder I have a folder that is my mods name - (exactly as it appears ... including .esp at the end)
In your mods folder (above) I then have a folder that is the exact name of the voice type I wish to use on my npc. So for me, the voice type file I created in the CK for my custom npc is called THJackVoiceType - and that is the name of this folder.
Inside this folder are all my voice files.
When you record inside the CK you should have a temp.wav file that self generates in the data/sound/voice folder. Hit record, record your dialogue, and when you stop this file should generate automatically.
Then if you click the flag "from wav" and hit save you should then get two files in the processing folder. A temp.lip and temp_resampled_audio.
Rename temp.lip to the file name generated in the edit response window in CK and put it into your voicetype folder along with another automatically generated voice wav that should already be in there.
Then copy the voice.wav, paste it in the same folder and rename it .fuz
You should now have 3 files in your voicetype folder
Personally I then go and delete the temp file in the voice folder and then the two files generated in the processing folder (not fonix) - to keep it all neat.
Once you have these files, in these correct folder paths they will play and show (lip) in game. To change the voice file, to something of far better quality, just name it the same as the voice file created with this process and replace the file you generated through the CK.
Do not forget to generate a new seq file using sseedit after doing any new dialogue or it will not show up in game.
Hope this helps a little, and again apologies if this does not.