9 ms·

I guess the next frontier for realism is character animation. To me their robotic movements and moonwalking feet break the illusion. There were very neat demos

by pornel 4y ago

I guess the next frontier for realism is character animation. To me their robotic movements and moonwalking feet break the illusion. There were very neat demos of ML-generated postures/movement, so it may be possible to nail it.

Eduard 4y ago
I agree. This game is in the cerebral palsy category of character animation (e.g. the NPC tourists on the canal's other side). GTA 3/4/5 are in the megalotesticle category (characters moving like wavering boats).
fabian2k 4y ago
There are games with very realistic character animation like e.g. The Last of Us Part 2. But that is a lot of effort, requires a lot of expertise and is likely very expensive. Correct animations can also conflict with fast and responsive movement, and most games will prefer responsiveness to realistic animations. Having both requires advanced approaches that can blend different animations well, which seems to be rather difficult.
- antegamisou 4y ago
  It's funny, I remember reading somewhere that research on Computer Graphics, rendering specifically, is 'solved' because photorealism had been accomplished. What had been omitted though was that efficient photorealistic rendering, as well as the visual elements that human perception instantly picks up on such as material(s) translucency, lighting/shadows, and as you've already mentioned animations (and the bigger challenge of facial expressions!) is what everyone's after. And what we have at the OP is indeed impressive work, which still looks artificial. Contrary to the critique of the comments so far, and outside the graphics research interpretation, this isn't something necessarily negative imo. I like video games that preserve a few elements of non-photorealism.
  - fooker 4y ago
    Human perception does not pick up bad lighting or shadows, this is a fairly well known fact in the computer graphics world. Sure, you can compare two and tell which one is better but unless it is really bad humans won’t be bothered by it or even explain what’s wrong.
- whoooooo123 4y ago
  This is an interesting video about how hard it is to effectively animate just one small aspect of human motion: walking through a door: https://youtu.be/AYEWsLdLmcc https://youtu.be/AYEWsLdLmcc
presentation 4y ago
Just played Yakuza: Like a Dragon and the NPC animation there as well as the animation and art style in general looked better in that game than in this clip.
sfvisser 4y ago
Lol, looks pretty realistic to me. This is very much like how most tourist in the city behave!
ufmace 4y ago
That's what I was thinking. Some people are complaining about very subtle details not being quite right. But to me, what really stands out as video-game-y is the way the character and camera and other people move. It's hard to describe, but they just don't move like actual people do. Gotta watch some bodycam videos for comparison. Don't get me wrong, the visual clarity is a great achievement. But if they're gonna spend more effort on making it more convincingly real, that's a better target IMO than reflections and wave patterns.
- Bakary 4y ago
  I'd be interested in hearing the perspective of a video game programmer, because from my own intuition as a layman, it sounds as though making realistic model movement is an order of magnitude harder than working on the environment alone. At least, the fact that this game has familiar human movement tropes that resembles decade old games sounds like a testament to that.
  - rhn_mk1 4y ago
    Not exactly general motion, but the linked video is about fluid motion in games: https://news.ycombinator.com/item?id=31200393 https://news.ycombinator.com/item?id=31200393
- cmehdy 4y ago
  For characters animations don't blend too well with each others as some things reset between postures, and the walking animation doesn't translate too well to crowded spaces as people would turn their bodies to maximize the distance between body and objects (up to a point), sometimes even opting to take a different path depending on people. And the camera is very smooth, which filters out the little balance shifts we all do (even typing something here I notice the subtle changes to vision oscillating in complex patterns which head-bobbing would be a simplified first-order approach to). I assume it's still very difficult to do all these things with current computing power, and in the case of the camera I'm not sure it will ever be possible for something interactive to feel quite perfect unless you switch to a VR headset and carry your own balance into it.
jayflux 4y ago
100% It feels like character movements haven’t changed much in the last 20 years. Which is a shame because the visuals are great, but let down by 2006 style animation.
- porknubbins 4y ago
  Yeah its to the point where you can instantly recognize someone doing a parody of a GTA character walk. I have no idea how it is implemented but it looks as if character limbs move at a smooth constant speed. As a runner I have seen graphs that show human locomotion is in fact made up of brief bursts of speed followed by a sharp decceleration. I’m sure they could model this but its probably not a major priority.

by pornel 4y ago

Eduard 4y ago

I agree. This game is in the cerebral palsy category of character animation (e.g. the NPC tourists on the canal's other side). GTA 3/4/5 are in the megalotesticle category (characters moving like wavering boats).

fabian2k 4y ago

There are games with very realistic character animation like e.g. The Last of Us Part 2. But that is a lot of effort, requires a lot of expertise and is likely very expensive. Correct animations can also conflict with fast and responsive movement, and most games will prefer responsiveness to realistic animations. Having both requires advanced approaches that can blend different animations well, which seems to be rather difficult.

antegamisou 4y ago
It's funny, I remember reading somewhere that research on Computer Graphics, rendering specifically, is 'solved' because photorealism had been accomplished. What had been omitted though was that efficient photorealistic rendering, as well as the visual elements that human perception instantly picks up on such as material(s) translucency, lighting/shadows, and as you've already mentioned animations (and the bigger challenge of facial expressions!) is what everyone's after. And what we have at the OP is indeed impressive work, which still looks artificial. Contrary to the critique of the comments so far, and outside the graphics research interpretation, this isn't something necessarily negative imo. I like video games that preserve a few elements of non-photorealism.
- fooker 4y ago
  Human perception does not pick up bad lighting or shadows, this is a fairly well known fact in the computer graphics world. Sure, you can compare two and tell which one is better but unless it is really bad humans won’t be bothered by it or even explain what’s wrong.
whoooooo123 4y ago
This is an interesting video about how hard it is to effectively animate just one small aspect of human motion: walking through a door: https://youtu.be/AYEWsLdLmcc https://youtu.be/AYEWsLdLmcc

presentation 4y ago

Just played Yakuza: Like a Dragon and the NPC animation there as well as the animation and art style in general looked better in that game than in this clip.

sfvisser 4y ago

Lol, looks pretty realistic to me. This is very much like how most tourist in the city behave!

ufmace 4y ago

That's what I was thinking. Some people are complaining about very subtle details not being quite right. But to me, what really stands out as video-game-y is the way the character and camera and other people move. It's hard to describe, but they just don't move like actual people do. Gotta watch some bodycam videos for comparison. Don't get me wrong, the visual clarity is a great achievement. But if they're gonna spend more effort on making it more convincingly real, that's a better target IMO than reflections and wave patterns.

Bakary 4y ago
I'd be interested in hearing the perspective of a video game programmer, because from my own intuition as a layman, it sounds as though making realistic model movement is an order of magnitude harder than working on the environment alone. At least, the fact that this game has familiar human movement tropes that resembles decade old games sounds like a testament to that.
- rhn_mk1 4y ago
  Not exactly general motion, but the linked video is about fluid motion in games: https://news.ycombinator.com/item?id=31200393 https://news.ycombinator.com/item?id=31200393
cmehdy 4y ago
For characters animations don't blend too well with each others as some things reset between postures, and the walking animation doesn't translate too well to crowded spaces as people would turn their bodies to maximize the distance between body and objects (up to a point), sometimes even opting to take a different path depending on people. And the camera is very smooth, which filters out the little balance shifts we all do (even typing something here I notice the subtle changes to vision oscillating in complex patterns which head-bobbing would be a simplified first-order approach to). I assume it's still very difficult to do all these things with current computing power, and in the case of the camera I'm not sure it will ever be possible for something interactive to feel quite perfect unless you switch to a VR headset and carry your own balance into it.

jayflux 4y ago

100% It feels like character movements haven’t changed much in the last 20 years. Which is a shame because the visuals are great, but let down by 2006 style animation.

porknubbins 4y ago
Yeah its to the point where you can instantly recognize someone doing a parody of a GTA character walk. I have no idea how it is implemented but it looks as if character limbs move at a smooth constant speed. As a runner I have seen graphs that show human locomotion is in fact made up of brief bursts of speed followed by a sharp decceleration. I’m sure they could model this but its probably not a major priority.