Recruitment for an ES4:Oblivion project

A place for Ogre users to discuss non-Ogre subjects with friends from the community.
Post Reply
MrTissueBox
Gnoblar
Posts: 2
Joined: Mon Jul 13, 2009 7:00 pm

Recruitment for an ES4:Oblivion project

Post by MrTissueBox »

I would post this in the recruitment subforum, but that seems to be OGRE projects only.

There is an Oblivion project going called the Oblivion Visual Enhancement Project, and many of us would like to see this project come to light for the game: The Elder Scrolls IV: Oblivion. What it plans to do is to make the graphics of Oblivion look better with true depth of field, better water visuals, etc.

The thread is here: http://www.bethsoft.com/bgsforums/index ... 93021&st=0

This is a list of what the ultimate plans were for the project:


Shadows-
[X] Summed-Area Variance Shadow Maps: Looks better and performs faster than percentage closer filtering, which is what we have in Oblivion right now. Win-win overall, and very possible given the data I have access to.
NOTES: Currently being reworked to use a 5x5 Gaussian blur and potential physically-plausible penumbra generation dpending on performance costs.
PAPER: http://developer.download.nvidia.com/pr ... /SAVSM.ppt

Clouds-
[ ] Simulated Volumetric Clouds: Oblivion does appear to have a separate shader for these already, contrary to what I thought initially. While I can't do *true* volumetric clouds per se I can fake it pretty well and even use some of the same algoritms.
PAPER: http://ofb.net/~niniane/clouds-jgt.pdf (one of many on the subject)

Water-
Note: as of 12/18/08, research is currently being done into having OBGE draw a separate water mesh using the screen-space projected grid algorithm.
[X] Parallax Mapping: Who needs dense meshes when you can do all that occlusion stuff in a pixel shader?
NOTES: Completed, pending view vector derivation.
PAPER: http://vrsj.t.u-tokyo.ac.jp/ic-at/ICAT2 ... /01205.pdf
[ ] Refractive Dispersion: Although plausibility is dubious at the moment unless I can find a way to create water refraction in a more... graceful, if you will, fashion, the effect is indeed pretty spiffy.
[ ] FFT water heightmap generation: A more realistic water simulation algorithm used in CG movies and Crysis. (On higher settings, at any rate.) I may combine this with a little bit of Perlin noise if the effect needs further tweaking.
[X] Cook-Torrance Specular Term: A physically-based shading algorithm that fixes the inaccuracies in the Blinn and Phong models regarding specular highlights at glancing angles. Much more expensive, but supports a lot of additional stuff like surface roughness and geometry shadowing terms. If I can recycle some of the Fresnel term stuff from the reflection bits this will be considerably cheaper.
NOTES: Completed, pending view vector derivation.
PAPER/PRESENTATION: http://www1.cs.columbia.edu/~belhumeur/ ... Models.ppt
[ ] Oren-Nayar Diffuse Term: Another, similar physically-based shading model for computer graphics, but accounts for diffuse shading instead. Also supports per-component indices of refraction and surface roughness.
PAPER/PRESENTATION: http://www1.cs.columbia.edu/~belhumeur/ ... Models.ppt
[X] Soft Water/Land Intersection: Little trick found in Crysis that smooths the transition between water surface and terrain. Although it may look odd in theory in practice it noticeably increases realism.
[ ] Procedural Underwater Caustics: You know, those little bendy patterns that show up everywhere in the pool, etc. Side effect of light refraction into the watahh.
DESCRIPTION: http://http.developer.nvidia.com/GPUGem ... _ch02.html

Skin-
[ ] Multipole Diffusion/Transmission: Craig Donner and Henrik Wann Jensen's physically-plausible algorithm for handling general subsurface scattering. Especially good at skin; used in the vast majority of high-end CG movies along the lines of LotR for doing just that. Note that I would need to roughly approximate this, however using a variant of translucent shadow maps the result would be fairly close.
PAPER: http://graphics.ucsd.edu/papers/layered/layered.pdf
[X] Cook-Torrance Specular Term: Now with settings ripped straight from nVidia's Human Head demo for maximum awesome.
NOTES: Completed, pending view vector derivation.
PRESENTATION: http://www1.cs.columbia.edu/~belhumeur/ ... Models.ppt

Hair-
[ ] Kajiya-Kay Anisotropic BRDF: Dump this texture lookup nonsense, make it use MATH! RAWR.
[X] Rim Lighting: Empirical technique for emulating some light backscattering in hair strands.
NOTES: Completed, pending view vector derivation.

Vegetation-
[X] Subsurface scattering: Evaluation of the Henyey-Greenstein phase function to simulate forward scattering. Pretty cheap, looks nice.
NOTES: Completed, pending view vector derivation.
[ ] Improved Alpha-Tested Magnification: Interesting technique Valve cooked up. This should improve the outlines of grass and foliage tremendously.
PAPER: http://valvesoftware.com/publications/2 ... cation.pdf

General/Terrain-
[X] Parallax Mapping: Regular parallax mappin' on teh terrainz.
NOTES: Completed, pending view vector derivation.
PAPER: http://vrsj.t.u-tokyo.ac.jp/ic-at/ICAT2 ... /01205.pdf
[X] Cook-Torrance Specular Term: Yeah, just an adaptation for more general use.
NOTES: Completed, pending view vector derivation.
PRESENTATION: http://www1.cs.columbia.edu/~belhumeur/ ... Models.ppt
[ ] Oren-Nayar Diffuse Term: Yup, same deal here.
PRESENTATION: http://www1.cs.columbia.edu/%7Ebelhumeu ... Models.ppt
[X] Ambient Aperture Lighting: Cheap method for applying ambient occlusion and heightmap-based self-shadowing to surfaces.
NOTES: Completed, pending custom art asset generation.
PAPER: http://www.cse.ust.hk/~psander/docs/aperture.pdf
[X] Distant LOD Ambient Aperture Lighting: The original application for AAL was on terrain. With some preprocessing, it would be very possible to add a very cheap shadowing effect on distant land. It might even be possible to implement a full area light for the sun depending on available data for the original DistantLOD shader.
NOTES: Completed, pending custom art asset generation.
PAPER: http://www.cse.ust.hk/~psander/docs/aperture.pdf

HDR and Tone Mapping-
[ ] Retina Burn: Made by Timeslip for his SM2.0 HDR implementation and ported over into the "real deal." Now without the funny afterimages caused by HUD elements.
[ ] Cinematographic Flares: Blobby bloom is boring. Yay for alliteration. Why not mix it up with some more interesting effects?
[X] Blue Shift: Emulates how the eye reacts biochemically to low-light environments.
NOTES: Bleach bypassing seems to make this go away almost outright. Bleh. I can probably change things to have the bypass applied within the HDR shader itself, getting rid of all the double-application idiocy.
[ ] Fixed-Point HDR Encoding: A special treat for all the bloom users out there. This would have to be done AFTER everything else is finished (as a rule!) BUT I would be able to give you fellas real HDR complete with the bright brights, dark darks, and everything in between. The catch is that it screws over alpha blending due to how it works. I don't think that would be a problem, though, as Oblivion generally seems to use alpha testing for a lot of stuff anyways and it all appears to be done in shaders to boot. (i.e. I can take care of it in there with no adverse visual impact) Did I mention it wouldn't break MSAA?


If any of you would like to help this project come to pass that would be wonderful and a great addition to Oblivion.
MrTissueBox
Gnoblar
Posts: 2
Joined: Mon Jul 13, 2009 7:00 pm

Re: Recruitment for an ES4:Oblivion project

Post by MrTissueBox »

So no takers? I love Oblivion, I just wish it would have better graphics.
User avatar
wacom
Gnome
Posts: 350
Joined: Sun Feb 10, 2008 2:07 pm

Re: Recruitment for an ES4:Oblivion project

Post by wacom »

MrTissueBox wrote:So no takers? I love Oblivion, I just wish it would have better graphics.
Haha, I think the exact opposite: Oblivion looks good, but I really can't love it. ;)
Post Reply