This article is kinda interesting and goes into some of the technical limitations around at the moment. It talks about how the introduction of the PlayStation eye and kinect has made it possible for smaller companies to create mo-cap. It also talks about creative decisions that the Enslaved team took when it came to merging the divide between cut scene and game play by adding in an extra camera which would be cut to at moments the developers thought were appropriate. Also because of the availability of mo-cap to them the work flow changed meaning they could add in extra little things they thought would add to the immersion as the process went on.
There are more parts to this article which are all really interesting to read but this section had more about animation processes
I found a forum where someone was asking about facial animation in games, limitations and work flow for their own thesis. It seems that the people responding in this forum have experience although how much is unknown so these responses could not be complete ones. Here's one of the responses
"Limitations are a key thing with games, in general. In fact it might be easier to ask what limitations they don't have.
As far as I know, most game engines are limited to bones, although some can also use blendshapes. Some engines are limited further by the number of joints they can use per character, the number of joints that can influence each vertex, and so on. More bones and weights requires more memory and processing. Many games will save resources by have separate models for game play and cinematics.
As far as animating the bones, sometimes you'll have controls similar to what's used elsewhere. Other times, you may have to animate the joints manually. One game I worked on even treated facial animation like music, or typing - specific expressions would be assigned to keys on the keyboard, and you would try to hit those keys at just the right moment while playing everything back, while a script recorded the keystrokes and translated it to bones."
The last part of this was particularly interesting. And this approach would explain a lot of the unnatural facial expressions in games currently.
"With games, you're a slave to the frame rate. Everything has to be optimised so that the game can run at a consistent rate, even in scenes with dozens or even hundreds of characters. The second the action starts to slow down, the developers will look for corners to cut, that don't affect the game play. Usually model detail, texture resolution, and extra bones are the first things to be sacrificed."
Another reason why Animation is often simplified.
Is low poly really a reason for limited facial animation? I don't think so. Ken Perlin's rig clearly demonstrates a large variety of facial expressions with very limited polys
Anticipation in game animation
I wanted to look into game animation while thinking about the principles of animation, one i was particularly interested in was anticipation as i felt it was one principle that generally is cut short. This sentence backs that theory up
" In games, an animator will be asked to strike a fine balance between instantaneous character response and believable movement. " (Mark Garabedian from animation mentor)
And added to by Sam Yip who works at Disney Interactive
"The anticipations in video game animations could be just a few frames,"
"if the frame rate is low and those few frames are skipped, then the player will not see the anticipation at all, which lessens the impact and believability of the animation."
DishonoredSide Note: Noticed an issue with dishonored (probably more about AI but i thought it was funny) A guard was stamping rats and then another came and joined him. A third appeared and decided it would be a good idea to shoot the rats killing both the rats and the two guards in the process. He then wandered off like nothing had happened. No animations for accidental friendly fire? Or interactions between NPCs? GTA4 had animations for when they bumped into each other but is that enough? I guess it's not that often that it would happen but would it make the game feel more alive if NPC's did interact more with one another. Like when they're having conversations with one another would it help if they didn't just stand next to each other and maybe gestured when they speak?
Test: Animate a gesture that people use when speaking, test if the gesture would be usable for multiple lines of dialogue. Can a few different gestures be enough to make it look convincing.
Evidence: Look at games where NPCs interact to see what (if any) gestures are already used, evaluate how effective they are.
Spec Ops: The line
Other Side note: James has been playing Spec Ops: the line for me today. It does seem to be more in line with the sort of Call of Duty and Battlefield sort of game play but as the review from Yahtzee and other sources suggested it has a lot more focus on how the story affects the characters psyche. He animations and actions get more extreme as the weight of his past choices play on his conscience. As James has played we have noticed that the way he attacks hand to hand has changed to be a little more extreme than when the character was calm and collected at the beginning of the game. An example is a new animation that been cropping up is as a take down move, beat an enemy to the floor and shove the nozzle of your gun into their mouth. A little excessive. His walk and run cycles have been affected by the injuries he's sustained as well.
If you want a game to make you feel completely horrible then this one is for you. I'm just watching James play and i feel guilty about some of the options.
Life Drawing
The past 2 weeks I've gone to the extra life drawing class on a Tuesday I went all through last year as well for a few reasons.
- Drawing from life is far better than drawing from photos
- It's helping me get better with proportions and weight distribution
- It's relaxing and a nice little constant for the week
Week 1.
A collection of 2-5 minute sketches, one moving sketch and two 20 minute sketches.
Week 2.
We hadn't got a model so we drew each other (I'm so sorry.)
We hadn't got a model so we drew each other (I'm so sorry.)
Reference:
Sam Yip and Mark Garabedian http://www.animationarena.com/video-game-animation.html
Forum comments http://forums.cgsociety.org/showthread.php?t=884223
Ken Perlin Facial Demo http://mrl.nyu.edu/~perlin/facedemo/
No comments:
Post a Comment