Hacker News Reader

Comment by zaroth | Hacker News Reader

So much for uncanny valley, to me that just looked fabulous.

I don’t quite comprehend the development process to build a short line this. Everything from the surroundings and environment, to the humanoid model and making natural movement, obviously the facial modeling and expressions, the water effects.

The sound track is probably the only thing I feel like I have a decent mental model of how it comes together.

I took one 3D graphics class in undergrad and it was all so absurdly low level basics, most just mathematics, it never remotely showed how something like this could actually be put together.

View on HN

I used to work for a game studio a long time ago. Here is a rough sketch of what the process looked like for that game studio

a. Concept artists will create multiple possible worlds by drawing either using a physical medium or on a tablet.

b. Art Director will critique concept art and iterate with concept artists until he finds a world he approves. Concept artists will then draw many scenes in this imaginary world to confirm that there is no miscommunication between the Art director and the concept artists.

c. Some of the key characters and scenery from this world will be given to Modelers who will create 3d models of those assets.

d. Initial goal will be to bring one or two scenes to life.

e. Once the initial few scenes and main characters are fully approved, more headcount will be added and the creation process becomes more parallel.

f. Everyone involved watches the short multiple times a day to monitor progress. There is accountability for who owns a particular character for a given frame.

I got the impression the parent was referring to the technical side of rendering such photo-realistic visuals real-time.

Ah, sorry misinterpreted the question.

Regardless, I found your description of design process & workflow extremely illuminating (no pun intended). Have not seen this succinct and informative descrption anywhere else.

Yep me 2. Thx dman

Blog posts that dissect the rendering of things like GTA 5 really show how much goes into making scenes like this. http://www.adriancourreges.com/blog/2015/11/02/gta-v-graphic...

Wouldn't part of the point of realtime raytracing would be to avoid a lot of this multipass trickery in the first place and just have a more sophisticated light transport model?

Could be wrong here, but I think most uses of raytracing like this limit it to shading and use traditional method for shadows. (Shading being the way that light "appears" on surfaces, shadows determining roughly how well illuminated a surface is. In theory you could use raytracing for both, but there may be limitations?)

I have no idea how RTX actually works, but based on the fact that exploits the existing rasterization architecture, I assume it adds at least one (very fast) pass.

Yes, that seems likely. But can you use it to drop some of the others?

Probably, yes. The only issue I see with that is the fact that Raytracing is only supplementing current rendering methods, not replacing them yet. The performance is not there yet. In a few generations we might see fully raytraced games, but not for a little while I would guess.

Realtime raytracing hooks into an existing multipass pipeline like this by replacing select traditional techniques. The easiest to replace are shadowmaps and AO. You can also make use of radiance (diffuse GI) and reflections with one or two bounces. Traditional techniques are currently faster and give better results for just about everything else. Also, you can only afford maybe 1-2 raytrace samples per pixel at the moment, even with RTX, so they make heavy use of AI denoising techniques.

>>I took one 3D graphics class in undergrad and it was all so absurdly low level basics, most just mathematics, it never remotely showed how something like this could actually be put together.

I feel like this is a huge problem in higher education especially: bottom-up learning leaves students bored, confused and demotivated because they don’t understand the purpose of what they are learning and have no immediate use for it.

Whereas top-down learning, where you start with the end result and work backwards/downwards from it, can provide the enjoyment and motivation necessary to persevere, especially when the subject matter is challenging and the road long and arduous.

They're gonna hold a "Making of Troll" tech talk this week, hopefully they release video from it.

> NVIDIA will be hosting a demo and “Making of Troll” tech talk with Epic’s Marcus Wassmer and Juan Cañada at this week’s GPU Technology Conference in San Jose.

https://www.unrealengine.com/en-US/blog/troll-showcases-unre...

> So much for uncanny valley

To me, the mouth always seems to give it away. Fabulous nonetheless though.

I liked the way mouth movement was captured in LA Noire. It was a bit low-res, but looked way more organic than any other technique I've seen. I'm guessing the production is complicated and the cost is prohibitory though.

Yeah, that's always been my issue too. There seems to be something about the elasticity of the human face, and the interactions of all the small muscles, that seems to be especially difficult to model.

This demo does at least seem to deal better with the darkness in the mouth. That has also always been problematic and made a lot of characters look positively demonic :)

Micro expressions are a pain to animate and yet every human we interact with has them pretty much, our brains are very effective in reading them as well since for our ancestors they where literally life and death.

Without them I think our brain rings all the alarm bells, a human face without them is odd and possibly threatening.

> making natural movement

To me, the movement didn't seem natural. But what it did seem like is classic Disney animation!

Reply on news.ycombinator.com