Talk:Faceposer reference

From Valve Developer Community
(Redirected from Talk:Faceposer)
Jump to: navigation, search

Mysterious Phoneme Data Loss

Faceposer is insanely buggy. I spent 2 1/2 hours working on a 17 second .wav file and saved it and reopened it like the many times before only to find out that ALL phonemes had been deleted. What the fuck. tundracool

Check if the "_work" file for your .wav still exists. Otherwise, you could export the word data as a fallback if something like this happens again. -- Sodabus 18:22, 26 Apr 2006 (PDT)

SAPI Language Code Errors

hi i want to lip sync so i load an WAV into it and then do redo extraction then i get this error: Error:SAPI 5.1 unable to create phoneme converter for language 1043 i have API5.1 intstalled.. any1 got an solution for this thx

What language are you trying to lipsync? ^Ben
Well i just want to do english its in faceposer so.. my OS is Dutch though dont know if thats interfere's Yeh I think it is, language 1043 is dutch.
Look here. It's untested, so beware. -- Sodabus 12:09, 6 Jul 2005 (PDT)
many thanks :D only my norman detects it as an malicious file, but thats always with norman so i guess its good :)

Further and future updates

Judging by the leap in facial technology between half life and ep1 anyone know if we're going to see an updated version of face poser too? Or will the current version be able to cope with the extra detail of (specifically) alyx's face?

It can't cope as it is. I will be suprised, shocked in fact, if the latest version of Faceposer isn't included in the Ep1 SDK update. --TomEdwards 08:41, 9 Jun 2006 (PDT)

Anyone out there?

Is there anyone else out there who choreographs? I've not seen one bit of serious choreography in any mod or map that I've played. --TomEdwards 08:45, 9 Jun 2006 (PDT)

I guess people are mainly focused on multiplayer, or the learning curve is too steep for most. -- Sodabus 11:18, 9 Jun 2006 (PDT)
I'm down to get to grips with some choreography some time later this month. To be honest, I find Faceposer pretty hard to get to grips with, largely because the documentation here isn't nearly as polished and complete as what we have on Hammer and entities. I've got the absolute basics down (MoveTo, LookAt, Speech events and Triggers), but beyond that I'm lost, even on stuff like ramping animations at the moment.
I dunno why I'm sharing this, aside from in hope that someone takes sympathy and starts writing choreography articles. Oh and my compile is taking longer than I expected. :) Perhaps you can find some choreographers in the machinima community? --Giles 11:54, 9 Jun 2006 (PDT)
There are tonnes of articles. And as for ramps: Open the tab, Ctrl+click to create a point and define one. --TomEdwards 11:58, 9 Jun 2006 (PDT)

I do a good amount of choroeo, but my buddy does it almost exclusively for our mod. the mod is The Battle of Philadelphia by the way (the wiki page is pretty out of date though). Vecima 08:47, 28 May 2008 (PDT)

Faseposer crashes


I'm doing the final project on my masters; it's the outtakes from G-man's Intro-speech: "The purpose of the production is to produce a "gamie" (movie produced with assets for a game) with the outtakes from filming of G-man's Intro speech from the Half Life 2. G-man is the (one of) the antagonists in the game and a recurring character from the first Halflife-game. The intro of Halflife 2 is G-man addressing the player directly as Gordon Freeman, the main character in the game and the player's avatar. G-man is arrogant and condescending. The problem is to create a set of outtakes that is humoristic and show the interaction between the actor playing G-man and the "director" of the movie".

So far I have: - Written the manuscript - Recorded the dialog - Edited the sound-files (adding samples from The Matrix & The Simpsons).

However, the editor crashed.

I do the following: the editor starts with the Gman model loaded - I can load the file into the phoneme editor - I press re-extract - I copy/paste the text from the manuscript - It manages to slice the file up, but then it crashes.

The bug is 100% reproducable.

Any ideas as to what is happening?


There is no known workaround until Valve start work again next week. You might want to try installing a more recent Microsoft Speech SDK, but that isn't a guaranteed fix. Phoneme data isn't strictly needed to create a scene though, so unless the deadline is Monday you'll be OK. --TomEdwards 14:43, 5 Aug 2006 (PDT)
well, the deadline is friday + I have to write a paper about it as well. So the alternative must be to load the .vaw-files straight into the choreography, right? But when I add a channel and pick ".vaw...", it opens up a library of audio-files from the game. I can load these, but I can't make it open another library, so that I can load the files I've made myself. How do I load my own audio-files onto the choreography?
I think you mean .wav. ;-) Unless you have added the WAVs to the soundscripts, a process that is perfectly acceptable to skip given how fiddly it is, you are going to have to reference them directly. They should be under <game>/sound/: just add the relative path to the Sound field from the Faceposer dialogue. So if you have a file at <game>/sound/gman_proj/gman1.wav, you would add gman_proj/gman1.wav. --TomEdwards 02:52, 6 Aug 2006 (PDT)

Phoneme Extraction Crashes Faceposer

I am in the difficult situation of using an HL2 mod as a stimulus for a research project. I've been running subjects using this HL2 mod stimulus for a number of weeks without problem. The interaction requires that I create new WAV files each evening before the participants for the study come in the next day. I record the participant names and use the phoneme editor to create the necessary mouth movement matching the name of the participant. Unfortunately for me, it appears the most recent update by Valve has completely screwed me in this process. Everytime I try a redo extraction, FacePoser crashes. If anyone is awake, please tell me you have a work around. Can I insert phoneme data into a WAV file without using FacePoser? If so, how? I need this info by 8/9/2006 9 AM PST. Your help is greatly appreciated!

Still no updates

Everyone seems to have fallen silent about this problem lately. I'm still getting crashed out with it... Is everyone else just using the work around or have i missed somthing. Also any idea who this is effecting? Does everyone have the problem, or is it just a select group of users? and if it is a select group... what have we got in common?

How to Auto Lip Sync

Im New to face poser, i need help to auto lip sync a voice file can any one help me?

QuickStartLipSync --TomEdwards 12:14, 20 Aug 2007 (PDT)

On startup crash - OrangeBox

I have recently encounted problems with faceposer crashing on start up. This only occurs if there is a character still open when the application is shutdown. After a few hours of banging my head against a wall, I managed to find a fix.

Open the hlfaceposer.exe directly from the bin directory not source SDK (eg. orange box "*\sourcesdk\bin\orangebox\bin" hl2 "*\sourcesdk\bin\ep1\bin")

Close the character(s) currently open (right click characters tab in the bottom right and choose close).

Shut down the application and relaunch from the SDK

Wave Crashes

I keep having crashes when using the Waves tab. It keeps happening. Here's what I am doing:

The .wav files are in sounds/vo/femvoice. I open up the Waves tab. When I expand the vo folder, hlfaceposer.exe stops working.

I need to load the Waves in order to make the scene. --Super200 16:15, 30 June 2011 (PDT)

  • Never mind, I solved it. As long as I don't expand the folder, Faceposer won't crash.