Video tutorial for Face Poser

Many people have asked me to make a video tutorial for faceposer, so I did. This doesn’t really get into heavy detail, but I’ll be adding more videos for appropriate things.
Tutorial 1: Lip Syncing

Tutorial 2: Choreography and Hammer;9819093;/fileinfo.html

Tutorial 3: Lipsynching through “Karaoke”;9819170;/fileinfo.html

Tutorial 4: Gesture Manipulation I;9819534;/fileinfo.html

Tutorial 5: Team Fortress 2;9819826;/fileinfo.html

Steam Community group: (for help, and open posting of Face Poser creations)

R.I.P. Stage6! All the videos are on Filefront now (using DivX). Please note that FileFront pesters me about deleting those files if I don’t notify that I want to keep them, so I would appreciate any downloads, to prove that they are not just collecting dust on their fileserver. I don’t mind if you put them on other video sites, just be sure you don’t take credit for it.

DISCLAIMER: I’m not familiar with the majority of errors one could get when using Face Poser. For the majority of reference, please consult the SDK forums at

One question that keeps coming up is when your WAV file is not quite recognized, and appears “blank” or its green line goes backwards when it’s added to choreography. This is most commonly because Face Poser doesn’t see the wav in the directory it’s looking in.
Its “assigned” directory is for whichever game you’re working on. If you’re modding Half-Life 2, it’s username/Half-Life 2/hl2, if you’re modding Episode Two, it’s username/Half-Life 2 Episode Two/ep2. To find out which game you’re “working on” you need to specify it in the Source SDK window before opening Face Poser. If you’re using the newest SDK, complete with the engine dropdown, you can just select from there. Otherwise, if you’re working on a game before the Orange Box, you will have to add the launch parameters: “-engine ep1” to the Source SDK launch menu. Right-click the SDK in Steam and go to Properties, then Set Launch Options to do so.
In summary: Just make sure the game you use is consistent. Are you making a scene in EPISODE 1? Then be sure to put the sounds there and start the SDK in there. Are you making a scene in vanilla HALF-LIFE 2? Etc.

Additionally, remember that unless you’re making a mod and actually knew what “soundscapes” were before you read this tutorial, then you don’t need to look for your sound in the big list shown (in the WAV file window) because it won’t be there. Valve and all the pros make script files which list details of each sound file in the game; for simple stuff, you don’t need to do that, and I personally see no benefit. What designates the sound used is the text area UNDERNEATH the large list.

If you are getting red errors on extraction, this may be for a few reasons.

  1. You didn’t install the Speech SDK from the Microsoft site (the download is marked SpeechSDK51.exe if I recall) (also, be sure you INSTALLED it, and didn’t just extract the winzip self extractor)
  2. You are using Windows Vista (it’s a problem with Vista’s version of the Speech SDK, or more accurately the way FP uses it, and currently nothing can be done)
  3. Your audio is choppy. If recording yourself, get a better mic. You can also take a few steps to improve your recordings: Enunciate. Don’t slur words. Hold the mic to the SIDE of your mouth to avoid wind on it. In dire situations, hire another voice actor. Finally, if taking audio from elsewhere, ensure there is no background noise. (in which case, consult tutorial 3.)

Hey thanks.

That was very good, thanks for that! But I want to watch your scene!

Will sure try that later.
Will notice you if I got problem :slight_smile:

Thanks a lot, you have cleared up some problems I’ve been having :D.

I can’t wait for the next tutorial.

Second one is up. Shows you all the basics of getting a scene running. No, I am not going to use setexpressionoverride because it breaks 3/4ths of the mechanics of Face Poser, and I’d really like people to learn Hammer anyway.

Agreed, Hammer should definitely play the major role with faceposing scenes.

Great work man. Also, ma-SHIN-ima. :stuck_out_tongue: Choreography has an h in it by the way.


Speech SDK 5.1:

For those of you who get extraction failure errors, click on Load then go to the HL2 directory and pick any scene file that’s already been created by Valve. Their scenes should work with extracting themselves.

Then, right click on the waveform and click on “Export word data as .txt” and save it with a filename of your choice. Then, load up your .wav file, but instead of re-doing the extraction, right-click on the waveform and click on “Import word data” and find the .txt file you just saved, then commit extraction. Bingo.

Be sure to delete everything except for one word and one phoneme inside of it. You can use that phoneme as your starting point and work from there. So now when you load up the Phoneme Editor, you can just click on import word data and you’ll be all set.

Using SDK faceposer in gmod is easier than making a map just to use a scene though.

Yes, and just changing mouth_drop in time to someone’s voice is easier than using Face Poser. Even Valve completely discourages any frequent use of setexpressionoverride; facial expressions don’t work well, you can’t have multiple actors, people can’t “look at !self” and will be glancing around at random stuff…ooh, a bird…anyway, I’m not going to say making maps in Hammer is easy, but it could lead to you finding a bunch of other ways you could make your scene more dynamic.

Great tutorial, thanks for posting.

Good job, and yes, Hammer is the key to a really well pulled off scene.

Thanks for the second tutorial :smiley:

I love you.

Thanks man :smiley:

Very good tutorials, the best on the net so far. Are you making any more?

Oh, thank god. The last FP tutorial I found, the guy had a Mexican accent.

I like your tutorial, also you have a nice voice.

Great tutorials there! Showed me a couple of things I was unaware of as far as faceposer goes. Would love to see more about Hammer, etc.

Interesting, but I’ll still use the gmod method until I discover some magic way of putting readymade maps into hammer. This tutorial would have been very useful to me a year ago, where were you when I really needed a faceposer tutorial?

At some point, I’ll actually be doing a tutorial on making scenes in existing maps.

Actually, why don’t I just list out the stuff I have prepared:

Karaoke (a method of lipsynching to low-quality audio)
More on Flex Animations
Multiple Actors
Adding a scene to a n existing location

Obviously, one method is to decompile then recompile the map. However, I might try to figure out a way to start it using ent_fire and ent_create commands.