Pendulum's Alter Ego face animation demo video


I like this Pendulum Alter Ego video of three different animated heads doing the same thing at exactly the same time. The hi-res version is here. Pendulum Alter Ego

31

  1. Aargh! Terrifying! I was expecting them to break out into a chorus of ‘Dead by dawn! Dead by dawn! Dead by dawn!’…

    Get it away from me…

  2. Are you sure the middle head isn’t the real one? If not, whoa, the valley is deep.

    Really? I saw this movie once… Final Fantasy: The Spirits Within, in 2001. ;)

    Now, granted that Square’s commercial failure in motion pictures opened the way for competitors to subsequently fill that role… such as, apparently, Studio Pendulum. And maybe it’s the crappy resolution Flash video of YouTube, but this doesn’t seem to me any more impressive than what was done seven years ago.

    Unless Alter Ego is software that I can run in real-time on my personal computer now. That would be impressive.

    Alter Ego’s Rocker is also impressive.

  3. i don’t think puppets, digital or otherwise, will ever be completely believable all on their own. it’s all about how and when they’re used, and unfortunately it’s rare to see them used well. “the spirits within” could have been far more compelling if it had used an overall aproach more like “the lord of the rings”, for instance: only use puppets where you have to, and film actual actors whenever you can.

  4. “the spirits within” could have been far more compelling if it had used an overall aproach more like “the lord of the rings”, for instance: only use puppets where you have to, and film actual actors whenever you can.

    Which is an interesting take on it since the thinking at the time by the people making The Spirits Within was the opposite — that not having to mix “special effects” in with live photography allowed for an entirely consistent viewing experience. In other words, if everything was “special effects” (but done so well that it seemed real to your senses), then you wouldn’t be “looking for the effects” anymore. It would all seem real.

    c.f. availability heuristic

  5. Why do the two men look so much more realistic than the woman? She looks botoxed to the max and way more cartoonish than the other two.

  6. Current the state of the art in digital actors creeps me out. It works in Shrek because it’s a cartoon(ish?), but when they do stuff like Christmas Train or FF it just looks goofy.

    It’s weird, the awkwardness of the movements of these characters give me the same sort of uncomfortable feeling I get when I’m watching Seinfeld or Curb Your Enthusiasm. It’s just.. weird. Uncanny one might say.

    I’m sure they’ll get there, but right now, yeesh, I’ll pass.

  7. I think the really interesting thing about this is how different each of the faces can look even while having the same expression. For instance the ferocious growl is much more effective when done by the young man than the woman. It says something about how much our features change how the people around us perceive our expressions, and how they react to us (or how we react to others).

    and also, it is really creepy! ;p

  8. “Why do the two men look so much more realistic than the woman? She looks botoxed to the max and way more cartoonish than the other two.”

    Wrinkles and crows feet?

  9. The crying was the least believable one for me…but the rest are past my uncanny valley. They didn’t creep me out at all.

  10. This was so eerie, because you’d never see faces coincide with the EXACT same timing in real life. Hypnotic, and I’d like to see more reactions, like maybe one with PhysX fluid and getting splashed in the face.

  11. @zuzu #9:
    i can understand their thinking on the approach, but i think they didn’t have a strong enough story to pull it off. what should have been a pretty-okay scifi flick is now a figurehead of bad 3D moviemaking. and addionally, the blending of digital effects with live-action footage has gotten so good that an “effect”, in the layman sense, could now be defined as anything impossible to film practically.

    when it comes to humans, it would seem, it will always be less effort and a higher reward to just film the real thing. that is, of course, when we’re just talking about filmmaking. in games, it’s a whole other thing.

  12. The facial expression is top-notch; I think the the unbelievable part of it is the skin reflections, not the animation.

  13. ZuZu @ 6:

    I disagree entirely :)

    I think the work here is far better than FF:Spirits Within, and I’m a massive fan of that movie.

    Spirits was definitely the landmark in animation for synthesizing photo-real humans and worlds, no one else had dared go there in a feature-length format, and it blew away everything at the time.

    But they overbanked on Mo-Cap, and got bogged down in data, which had to be massively edited/ de-noised/ down-sampled (typical of the time, somewhat less-so now), which led (or contributed substantially, at least) to a fairly understated acting style all ’round. Also the facial capture wasn’t that advanced.. even though they produced amazing results by, again, working the data by hand.

    Compared to that amount of post-process, the demo above is just amazing!, the stuff shown in the “demo”-type spots on the AlterEgo site show work that is virtually raw compared to the FF (and most other) Mo-Cap. It’s lights years ahead in terms of ease of use and stage-to-screen time.

    FF was very ambitious at the time, and what they achieved in their skin-shading and subsurface-scattering (or at least a hack of it) was phenonenal.

    But looking at the shots above, for me, is entirely different. The very subtle movements in the neck/shoulders cause the heads to really play about the spinal axis in a very human way. Spirits Within is bit stiffer than this, and definitely lacks the quick, real-life snappiness of the demo above. Spirits is certainly fluid, but it’s a little viscous compared to this.

    Where FF really excelled (skin/eyes/texture) back in 2001 is almost bread-and-butter here (not really, but check the studio above’s demo-reel). Importantly too, the subtlety and finesse of the facial expressions is an order of magnitude better than the face-work in FF. Even things like the eyelid/sub-eye-crease animation are noticeably better than so much stuff out there (sounds silly, but it’s true, and very important to believable work).
    _

    Also, and judging by the responses above this is probably going to be an unpopular view (not in the negative sense) but, for me, this has gently stepped to the side of the Uncanny Valley, and walked right on. I find it very easy to interpret the characters above as “human”, without getting the UV creep.

    Immediately upon watching it I thought “wow, these guys have cracked humans”. I genuinely don’t get the UV-creepy-hyper-real vibe off these characters (maybe at the very beginning, but when the characters show some variation of expression, it just falls away).

    I also wonder if people are possibly using the phrase Uncanny Valley in a different sense than I do.

    1. The only thing that fails for me, besides the fact that the males look human and the female looks like a high-end sex doll, is the teeth. They’re what are known in the makeover criticism business as chiclet teeth – way too regular and even. On a more personal note, if they print out the guy on the right in 3-D, I’d be willing to explore his uncanny valley.

  14. I found both males to be quite believable. But the female looked like a real-time in game rendering. Nothing like the other two.

  15. when it comes to humans, it would seem, it will always be less effort and a higher reward to just film the real thing.

    Not if “the real thing” doesn’t want to do the thing you want to film.

    Filmed confessions, faked-incitement, fake video-‘evidence’ of a crime.

    Give it 15 years, and we will have to actively suspect every video, of any importance, of being a cg-staged event, the same as we do for photos these days.

    I know we can do similar things with stand-ins today, but it’s going to be a very different game when the technology to get anyone to say or do anything becomes available to anyone who wants it.

    I’ve seen RunningMan.. ignore at your peril ! :D

  16. I’d be willing to explore his uncanny valley.

    Antinous !

    -1, that was awful !
    And you know what a fan of inuendo I am :)

    OW!

  17. Ant, how the magic-fun did that happen?

    I quoted you before you said it ! !

    Gremlins or trickery?!

    I am the future!

    1. I have my own little corner of the uncanny universe. My comments appear early and fall into queue late.

  18. Antinous and Arkizzle, we’ve warned you to stop playing with the entropy arrow! Get back into alignment with the timestream now.

    Seriously, though, the young guy’s teeth and mouth look too freaky weird for me. The old guy was very natural to me, and apparently the woman was cartoonish enough that she didn’t bother me, exactly. But didn’t look live, either.

    But the old guy — wow. (Except the eyes; the eyes were a little too Anthony Hopkins to seem natural, if that explanation isn’t entirely self-contradictory.)

  19. I find these expressions very good, but obviously a way to go to be convincing as humans.

    It’s like the faces are missing expressive muscles. In particular, the foreheads and the space between the upper lip and nose don’t seem to have independant motion at all. I think that in order to truly get it right the animators have to realize that expression is not just in the face, it’s the whole head. IF you watch a human smile, it’s pretty rare to not see motion in the ears and the rest of the scalp.

    More assymetry in the expressions would help, too.

    Also, I agree that these characters remind me of Shrek.

Comments are closed.