Anyone Using FaceFX?

Hey everyone!

This is my first post on here although I’ve been a long time follower of TAO. In all of my browsing though, I’ve never come across anything related to using FaceFX (http://www.facefx.com/). So far, I’ve done my best to create a working pipeline from 3DS Max to Unity, but there are a lot of aspects of FaceFX that I am still working to grasp. The company I work for is relatively small, and I am in charge of the pipeline and setup of characters using FaceFX. My major concern at the moment is achieving more than just lip-syncing the characters, but also using FaceFX to procedurally build in hand gestures and body motion. After working on this for a few weeks now, I’ve managed to achieve acceptable results, but I really don’t know if I’m going down the right path with my setup. There seems to be a distinct lack of information regarding FaceFX on the internet, with the exception of some OC3 tutorials and a few others that consist of mostly lip-sync setup. If anyone out there has had some exposure to this software or a situation similar to this, I would love to hear how you approached the problem and what kinds of issues and solutions came up.
Unfortunately, I can’t really go into great detail about the project because of NDA nonsense. I would be happy to share what I’ve learned so far about FaceFX if anyone is interested though.

Thanks in advance!

We use FaceFX for all of the facial animation in our game - and it works great. Several years ago we thought about using it to to control full body gestures as well, but we decided against it in favor of letting our cinematic designers select the gestures that are appropriate for each line of dialog. This solution works better for us, but you can probably get good results letting FaceFX drive everything as well.

Thanks for the reply, Ben!
Ultimately, it would be awesome if we had the time and manpower to create all the individual gestures and queue them selectively per line of dialog. The team is rather small and we have an incredibly tight deadline, so I don’t think this is going to happen. I’ll keep hammering away at the node graphs and analysis actors until I get some better results. After everything is said and done, I’ll probably put together some kind of small page describing the process I went through for new people to maybe learn a bit and also for some critiques from you veteran guys (hopefully!). Thanks again for the response!

Just don’t expect perfect results and understand that the costs of customization to procedural results are generally very expensive. This is generally true of any procedural system (including stuff like mocap). If you are on a small team and want to do this stuff procedurally using FaceFX, you definitely can, just know your limitations (either commit to building your own procedural system, and you’ll get better results but at a MUCH higher cost, and just accept the quality any procedural solution is going to give you).

Sorry, no practical technical advice.

Yep, the only feasible approach for us is to rely on procedurally generated content. It is definitely not perfect in all cases, but it is producing pretty good results so far. I’m sure it is only mediocre at the moment simply because I have not tweaked it enough yet. We’ll see what happens as it gets further along though. Thanks for the input :slight_smile:

It will never be great- you should probably aim for ‘slightly better than mediocre’ for the best bang for your buck. The more varied your outputs, the less you can do by reworking algorithms and tweaking.

In the next release of FaceFX (FaceFX 2013 due out closer to the end of this year), we have greatly improved the FBX pipeline (and therefore the pipeline with Unity).

Procedural body gestures is an active area of research, and a difficult problem. At this point, it’s more something that FaceFX can help you solve than solve for you. I’m hoping to make some progress on this front for the 2013 release using mocapped body animations as a starting point.