Fallout 4 0.288

Forum rules
new topics are not allowed in this subsection, only replies.
  • Author
  • Message
Offline
User avatar
*sensei*
Posts: 267
Joined: 12 Oct 2012, 19:14

Re: Fallout 4 0.288

Portal2x2 wrote:
ENBSeries wrote:Dude, it's not ssao.
How else to still get the depth but not through a normal? In the code all calculation of the occlusion goes through it, I already threw the source code, I recommend to get rid of the shift on the course, other assistance I did not hear, only loud statements that "this is not ssao" yelled golosino. I don't how don't doubt your knowledge, but be advised that it is possible, I'm no expert in hlsl, doing what I could learn, the code is almost not changed. Attach again. http://pastebin.ru/LnjHmOi7 Does it need to rewrite all of this, I was watching different tutorials and nothing contradictory is not found, it all fits, but again I'm no expert but an Amateur. Call it cell shading graphic in General barbarism. What you see in the screenshot only raw ssao is derived from the normal without the use of masks and ambient, diffuse only.
its not ssao. If u have ripped some AO code from other software and pinned it into enb, that doesnt mean that u will get an AO effect in game.
Proper AO never shows mesh topology. If it does - then u have wrong source data input on your shader. By wrong data i mean, that numeric space of depth is wrong.
And yes, in order to retrieve normals in enb you have to restore them from depth. Yet its very inaccurate without view space matrix to object space. Bcos u have no clue of Y coord basis positioning.
_________________
Image

Offline
User avatar
*blah-blah-blah maniac*
Posts: 619
Joined: 29 Dec 2011, 20:05
Location: Санкт-Петербург

Re: Fallout 4 0.288

AgainstAllAutority wrote: its not ssao. If u have ripped some AO code from other software and pinned it into enb, that doesnt mean that u will get an AO effect in game.
Proper AO never shows mesh topology. If it does - then u have wrong source data input on your shader. By wrong data i mean, that numeric space of depth is wrong.
And yes, in order to retrieve normals in enb you have to restore them from depth. Yet its very inaccurate without view space matrix to object space. Bcos u have no clue of Y coord basis positioning.
I did not take pieces of code, and took the entire source code, I posted it just above, could would like to check it out, though to open. Judging by the articles in this case I'm using is not the topology, and the retopologiâ meshes, this helps the effect to look for space darkening, and not just duplicate the mask based on the blur, and you must not forget that the ENB is limited to the bindings with 2D samplers, which creates additional problems when working with code. https://hejiraproject.files.wordpress.c ... xtures.png The right image, about it I try to use in the game. Here's another example of occlusion on fractal surfaces http://www.ray-tracing.ru/upload/articl ... les338.jpg Here is an example article http://ray-tracing.ru/articles232.html Yes the material on the network a lot, the algorithm is similar, already struggling with this code the second day, well, you can't differently use, only via normal, search for the game shadow of costly resources. Once again, I apologize but I'm looking into, and to me your like to the moon)) A huge plus of this approach is the improvement visible in space normals, regardless of 2D textures, even if she did not have time to draw, the geometry will be much better with no visible loss of frame rate. Don't know why but I have this code seemed to be the most that ninaest quality, and does not weigh a lot, look for something more difficult, writing from scratch is very difficult. Okay, let's take the old crysis and look at his ssao http://www.radioactive-software.com/urb ... s_SSAO.jpg the only difference is the dimming effect, because the calculation of the shadows, and out of my normal. While retopologiâ there.
_________________

AMD RYZEN 2600
Patriot DDR4 signature premium 16gb 3000ghz
MSI A320M-A PRO
Patriot burst elite SSD 480gb
hitachi HDS721050 500gb 7200rpm
Gigabyte GTX 1060 6gb

Offline
User avatar
*blah-blah-blah maniac*
Posts: 17551
Joined: 27 Dec 2011, 08:53
Location: Rather not to say

Re: Fallout 4 0.288

Portal2x2
Try at least increase sampling range, probably helps a little.
_________________
i9-9900k, 64Gb RAM, RTX 3060 12Gb, Win7

Offline
User avatar
Posts: 84
Joined: 21 Nov 2015, 14:15

Re: Fallout 4 0.288

Yeah it's just too ... thin. Which make the whole effect look much more like Cel shading than SSAO. I don't know if you improved the shader since the first time you posted in the screenshots section, but I gave you a few advices there, notably about depth comparison. In the exterior screenshots, the effect was drawn upon the background - the sky ! - behind trees, which is completely, utterly impossible with a correct SSAO code.
Also, in the power armor shots above, sometimes a surface will be darkened, but not the other surface that is obviously the cause of the darkening. In the end, it looks much like an incidence shader ( also named facing ratio ) which basically detects whether a surface normal is perpendicular or parallel to the camera and darkens it accordingly.


Which reminds me, is there a texturemap available with ENB that actually give us the whole scene normals ? We have Depth and Color, if we could get the normals rather than have to derive them from Depth maybe it could help a few effects ?

Offline
*blah-blah-blah maniac*
Posts: 565
Joined: 05 Apr 2014, 10:29
Location: Taiwan

Re: Fallout 4 0.288

AgainstAllAutority wrote: And yes, in order to retrieve normals in enb you have to restore them from depth. Yet its very inaccurate without view space matrix to object space. Bcos u have no clue of Y coord basis positioning.
Most SSAO solutions ignore Y/X, bias exist but most likely irrelevant.

Portal2x2,
take a look at AO papers/ sliders/ presentations. or try to read Marty Macfly's port at Reshade.
_________________
Intel Xeon L5639 6C12T @3.96GHz | Gigabyte ga-x58a-ud3r | MSI GTX680 4G | 48G RAM | Intel 760p Nvme w clover bootloader
Flickr
YouTube

Offline
User avatar
Posts: 27
Joined: 20 Jul 2015, 22:06

Re: Fallout 4 0.288

ENBSeries wrote:kranazoli
Not all displays have 60 Hz refresh rate, game will run like crazy from vsync enabled for 144 Hz display, so fps limiter helps here.
ShadowBoost have lower desired fps because 5% is dead zone where shadow distance is stuck, doesn't matter if it too short or too long, that's why users with stable 60 fps must have 60*0.95 or below frame rate set to force shadow distance computation if required.
I see and thank you very much for the detailed explanation.
_________________
AsRock Z77 Extreme4 | Intel Core i7 3770K@4,4Ghz | Kingston HyperX Fury Blue 32GB 1600MHz DDR3 | MSI GTX 980 GAMING 4G | Kingston 352GB SSD | XFX XTR Series 650W | 24MP57VQ-P IPS 24" | Windows 7 Ultimate SP1 x64

Offline
User avatar
*sensei*
Posts: 267
Joined: 12 Oct 2012, 19:14

Re: Fallout 4 0.288

kingeric1992 wrote:
AgainstAllAutority wrote: And yes, in order to retrieve normals in enb you have to restore them from depth. Yet its very inaccurate without view space matrix to object space. Bcos u have no clue of Y coord basis positioning.
Most SSAO solutions ignore Y/X, bias exist but most likely irrelevant.

Portal2x2,
take a look at AO papers/ sliders/ presentations. or try to read Marty Macfly's port at Reshade.
just an example of restored normals:
Image
camera pos1

Image
camera pos2

look at green component at structure's rooftop (same story with red component and left/right camera movement or rotation). Its changed. However object and surface still have same normal, it should be independant if camera coords. Situation gets worse of objects are closer to camera. This is reason why using any proper AO without normals is pointless.

And every proper AO technique requires normal vector, for proper hemisphere choice.

yet, classic AO dont need it indeed.



enbseries, I am not afraid to say that we are all waiting for your sweet AO inmplementation for fallout4 :roll:
Last edited by AgainstAllAutority on 21 Dec 2015, 15:39, edited 1 time in total.
_________________
Image

Offline
User avatar
*blah-blah-blah maniac*
Posts: 17551
Joined: 27 Dec 2011, 08:53
Location: Rather not to say

Re: Fallout 4 0.288

That buggy ssao reminds me bugs from native perspective correction of depth values on surfaces closer to parallel to camera direction. To avoid them trasforming to world space positions helps or increasing sampling range with clamp of minimal sampling range to visually reduce bug as much as possible. Detriangulation based on angles also do the trick partially. Anyway, copy-paste process for ssao is not good idea, user will stuck with depth to radius factors and sometime with wrong direction of normals. Understanding how it works is the only proper way and it brings many ideas how to tweak quality, improve performance, etc.
_________________
i9-9900k, 64Gb RAM, RTX 3060 12Gb, Win7

Offline
User avatar
*blah-blah-blah maniac*
Posts: 17551
Joined: 27 Dec 2011, 08:53
Location: Rather not to say

Re: Fallout 4 0.288

Folks, i wanna warn you about game bug which must be avoided at any costs while playing. The longer time you have spent in the game, the bigger error is because of absolute time counter. Increasing game time and then saving game is one of the reason to get bugs. Internally it's a float value of game days, which is the only (as i see) available in scripts, so floating point error increases according to the time passed. If you have more than 10000 in-game days passed, expect different behavior of the game. With about 100000 days it broken for 60 fps gameplay and require new game to be started or saved game to be patched to reduce time spent value. Values sounds like huge, but what i have seen in Skyrim statistics prooves that this is possible for game fans and i don't think they will be happy to loose very long time played save game.
_________________
i9-9900k, 64Gb RAM, RTX 3060 12Gb, Win7

Offline
User avatar
*sensei*
Posts: 316
Joined: 12 Aug 2013, 18:55
Location: Scotland

Re: Fallout 4 0.288

ENBSeries wrote:Folks, i wanna warn you about game bug which must be avoided at any costs while playing. The longer time you have spent in the game, the bigger error is because of absolute time counter. Increasing game time and then saving game is one of the reason to get bugs. Internally it's a float value of game days, which is the only (as i see) available in scripts, so floating point error increases according to the time passed. If you have more than 10000 in-game days passed, expect different behavior of the game. With about 100000 days it broken for 60 fps gameplay and require new game to be started or saved game to be patched to reduce time spent value. Values sounds like huge, but what i have seen in Skyrim statistics prooves that this is possible for game fans and i don't think they will be happy to loose very long time played save game.
Is the counter that GameDaysPassed global variable? If so, shouldn't the console command:

Code: Select all

set GameDaysPassed To 10
remedy the issue? Or does it go deeper than that?
_________________
Intel i7 6700k | AMD Vega 56 8GB | 2x16GB DDR4 @ 3000mhz | Windows 7 64bit | Creative Soundblaster X-Fi Titanium Fatal1ty Pro | Asus z170 Pro Gaming
Post Reply