Meet the Devs: The Heretic short film -- Post Your Questions

The register link goes directly to the zoom meeting.

 

For something that is constantly changing, like the bird, are you bending those lines manually between each frame?
Or, do you set some kind of wiggle protocol to waver within certain constraints and then see if you like how it turned out?

 

for people that would like to explore the implementation of alembic files, and unity as the main render tool in a pipeline what would you think is the best workflow that consern things created outside unity, specially characters, what would bet he maps the rigging techniques or how would be the most gratefull workflow to create quility short like "The heretic" for somebody just starting in this use of unity for film making, Thank you so much and i love your work

 

Thanks for a great presentation and some interesting insights, specifically into the Nested Timeline workflows!


Repeating my questions about Dirty Details here since there wasn't time to reply:

1. How many bugs (Core/HDRP) did you find throughout development?
   
   
2. Were you able to stay on "vanilla" HDRP?
   
   
3. Have you tried exporting to other platforms (e.g. VR)?
   
   
4. Is Heretic now used as a test case for e.g. Engine/HDRP updates internally?
   (this was answered ambiguously - if some of the internal scripts are not working with the latest HDRP release the answer seems to be "only partially"?)

 

Is it the imperfect photogrametry input that causes an area on the top left of the actor's head to flex in and out the early render of the actor's head?
It was on the right side of the picture early in the video showing the actor's face rendered in gray. It looks like he has a membrane on his head.

 

Besides a method for player input, what other work would still need to be done to make this into a game level?
ex: If I tried to open this as a game, would everything fall through the floor?

 

You did an awesome job and I love Unity and where it's ging. There are, however, two problems that I just couldn't find an answer for:
1. Are there any plans for developing realistic hair for Unity? Or at least something close to the current AAA title standards? There were some fine solutions that were discontinued, but sadly with no replacements.
2. What happened with Jim Hugunin's cloth simulation? It looked even better than in Marvelous and had a great performance in Unity 3 years ago.

 

There is a place when corridor changes and player enter the cave.

1. How did you do this? Was it used physics ?
   
2. Could you add support for Unity default destruction physics?

 

Congratulations to the team on a really amazing project! I just opened the Facial animation sample scene and my first impression is you've passed the uncanny valley in a spectacular way - to the extent that looking at him moving makes me a bit nervous because he doesn't blink

So the one thing I find really discouraging when starting a Unity HDRP project is that as soon as the sample scene pops up the first thing I notice is some really annoying shadow artifacts without even moving the camera. Since lighting is such a beast to handle and dynamic shadows are such an important part of graphics fidelity could I ask for any tips as to how to start going about setting up lighting, particularly the shadow cascades, so as to avoid noticeable transitions between higher/lower resolution shadows when approaching objects? How did you approach the lighting setup in The Heretic?

 

@lassejon Awesome! Looking forward to it!

 

Thank you

 

Fantastic!

 

Quick Question... wasn't the "Meet the Devs: Deep Dive into The Heretic Assets" webinar supposed to happened this past Wednesday? Is the one that is now on June 3rd... I wasn't able to attend the webinar that was on Wednesday and just want to make sure it's the same.

 

We only see a part of it, but to me, probably the most intriguing thing in this demo is the part where he takes off his jacket. The donning and removal of clothing has always been difficult to represent in 3D graphics, which is why it's generally avoided. At least as far back as Final Fantasy 8, game developers have been pulling tricks like fading to black and reskinning a character's clothing, or more recently just having it happen off camera while you're looking at something else going on nearby.

But in this, we watched the guy take off his jacket. Or did we? From the camera angle, we only saw a part of it, and not the interesting part (pulling his arms out of the sleeves). But it's still more than I've seen in games. Was this just the latest trick, or did you guys get an actual, working solution to clothing removal?

 

I just have one question... how in the world do have giant vistas like the city at the end? are we seeing actual geometry to scale with no clipping distance on the camera or is there some kind of 3d skybox / scale illusion going on? seems like there are are thousands of unique rock / building shapes and textures visible in that view. I can't wrap my head around it.

 

Okay but will we be able to do natural african type hair? I don't mean exotic style big curl like unchatred 4 or beyond good and evil, I mean really tight small curl.
I have been studying the issue and there is no paper, I find one attempts by Marschner BUT he was doing all wrong due to not understand the hair at all (he observe the hair is rougher and attribute that to the oval cross section, but his measuring

I had also to devise a new classification to describe this type of hair beyond the Andre Walker system (the FIA system is not better). Basically I have further elements for type 4 hair, like stress (horizontal or longitudinal) and diffuseness (curl staying together or not ), and observe that many people aren't aware of some aspects that impact visual ("baby hair" edge) or "spiral perversion".

Also that when the hair is flatten, there is a difference of texture with detangle hair and "ironed", where the micro structure are crumpled physically (or chemically), the same way you can't bring a slinky flat, it become cringly and have orientation all other the place (ie rougher as the normal relative to an area is all over the place), while in detangle hair, the highlight are a bit "wavy" (dotted regularly) because the structure is simply elongated. In my observation oval cross section make the hair finer and more fragile, visually there is a slight iridescence.

For physics, the closest approximation is the central core system pixar did for brave for much larger curl, but idea can be salvage for real time game with minor modification of existing algorithm. But another big factor is "lockness" as curl can get entangled in each other due to the 3d nature, but that can be approximated like we already do with flat hair by interpolating around a control curve.

It's important to note that spiral perversion have a huge influence in the physics (cause "shrinkage" by having the curl collapse into itself through overlap), twisty trajectory and creating locking in complex diffuseness (fun fact spiral perversion are created through mechanical longitudinal stress, combing that type of hair actually messier, that's why cornrow are used to protect the hair instead). Also african textured hair being more fragile mean that those parameter evolve along the length (tip tend to be looser due to continuous stress), braid and headwrap are ways to keep it healthy from stress. It's worth noting that horizontal stress, which push the 3d helix/spiral perversion toward a plane, is massively important in the visual, especially for silhouette.

I have designed a way to do the traditional card method for representing all these feature to keep a usual game workflow too. If you can address these aspects, you'll be ahead of any engine/paper already and it just mean small adjustment with the state of art. I'm working on another method, that would provide a core mass as background, for cheap hardware by tracing modulo space (without using the limited modulo operator, by making a mathematical extension of the DDA algorithm to intercept a moving boundaries that is the regularly space primitive, though I struggle with interval as my math is rusty).

I also want to address the skin shader. In my game, due to the setting being caribbean, I have been experimenting with skin shader applied to black character that are dark (i'm not dark skin, nor I have african type hair), but they all seems wrong in a way that match observation of AAA games by bleeding too much or not enough. I have reviewed the publications around skin shading (especially Jensen's work like spectral brdf for skin shading) and they all mention, when measuring the skin subsurface scattering, that black skin diffusion profile is actually different from white skin (due to the presence of mix of eumelanin and pheomelanin) but in game it seems we don't take that into account, and it make black character appear weird and downright white when the light is too strong (watchdog2 and I consider him brownish dark, when game don't actively avoid to make dark skin character, uncharted 4 isn't really good)

.While many twitter drama... I mean debate, seems to point at many people believing that's a common behavior for dark skin, my actual observation and sourcing (i'm in the caraibe) don't match that belief, there is enough source with instagram now too. I was wondering if the next demo could take that into account as it seems that it's documented enough and the subsurface parameters just need adjustment (potentially a skin slider to adjust ratio of melanin type). While there is a whole artistic tangent to do about lighting a dark skin (keeping readability with specularity dominating instead of shadow), it's out of the scope of technical adjustment.

Extra credit: Also I was wondering is there is a way to "bake" the subsurface scattering profile that unity preview in the editor into a ramp stack texture like penner's does with it's pre integrated skin shading? Or at least do you have any hint to where I should look in the code and how to do that? That would be massively cool!

 

This is probably a preview of what unity is cooking
http://unitytaiwan.blogspot.com/

 

Hey, thank you for sharing such awesome work and having the zoom session!

Is the Snappers rig used in the final film, or is that all 4D (vertex cache basically?) made with RussianScanner3d?

Also, i cant access the Unity.DemoTeam package from my scripts even though the full demo is running. Rider can find it but Unity cant when it compiles.

 

Can we use the Snappers system on our own rigged character?

 

Still would love to get an answer to my question.... thanks!

 

Where can I download this project?!

 

The register link goes directly to the zoom meeting.



Kodi