TES Skyrim 0.244

[jrista]

kaffeekranz
Okay, these are proper result, ssao/ssil work twice faster.

Oyama
Noise decreased a little for more or less planar areas, because new code using adaptive level of detail, but not a big difference. I'll try to optimize a little more, but this is not that optimization we spoke about, don't want to implement it, better to move to gi and keep current ssao/ssil as old effects.

:O :O :O

Were getting Global Illumination for Skyrim?? *DROOL* I really CAN'T WAIT!!

[spadesage]

Question: how in the bajeezus do I get the time-of-day to register correctly? As in, what numbers/ checkboxes in the TimeOfDay/ NightDay sections need to be present to get it to show the correct lighting for the time of day (and are there other things that need to be set up as well)?
I'm a complete "noob" at ENB, but I am quite familiar with coding and the like, so I sorta get how it works overall. It's just irritating that at 10:30 AM game-time it's pitch-black outside and the crickets are a-chirping (it almost seems to behave lighting-wise in reverse- as though night is daytime and day is nighttime). o.O

For reference, the current settings I have for Time of Day & Night/Day are as follows:
TIMEOFDAY: Enable=true DawnDuration=1.3 SunriseTime=6.55 DayTime=13.0 SunsetTime=18 DuskDuration=2.0 NightTime=1.0
NIGHTDAY: DetectorOldVersion=true DetectorDefaultDay=false DetectorLevelDay=0.65 DetectorLevelNight=0.25 DetectorLevelCurve=1.0
(Those are the settings for MACHIIMACHII's Saraan Suum ENB, but I'm using them with the Poupouri/ Morning Star ENB: both are by the same person-- I like the visual effect of 'Morning Star', but wanted the ToD effect from the other. I wanted to change the NightTime to 22, but saw no difference at all visually, so I set it back to what it was originally.)

I did manage to get a kind-of-but-not daytime effect by changing the "NightTime" number to 10.3-ish... which makes little sense to me (given that the max. number for each setting is 24, so I thought the thing used "military time"... and figured 22 would be an ideal number for the onset of "true" night).

So... what do I need to do to get the time-of-day to behave correctly? And would changing the time ratio (default for Skyrim is 10 game minutes per real-time minute; mine's set to 5 game minutes per real-time minute) affect how this works?
...Or is it just easier to try and find another preset that works and is more up-to-date (my current one is no earlier than November 2013- did it need to be even newer to work?)? And if so, where would I find one? 'Cause I can't seem to find any here or on the Nexus that are regularly updated. :S

[ENBSeries]

spadesage
I don't understand what is wrong and what is it "bajeezus". First of all, make sure you have helper plugin which guarantee precise time reading (download it from http://dev-c.com). Second, official properties for time of the day are set according to sun position on the sky and i don't recommend to modify them, but in your case they are not far from mine, so don't think it's problem.

Oyama
Almost finished ssao which you asked in pm. As expected it have about 3-4 times better performance, but it lowest quality have same performance as current middle quality because of additional computations required. SizeScale above 1 useless, better to add parameter which disable ssao filter when SizeScale=1. Not sure if this will be new ssao type or not implemented at all, i just show you comparison screenshots, is it really important or better to move to absolutely new instead of optimizing.

[Maeldun]

spadesage
Default timescale for Skyrim is 20, not 10. This means that one minute of game-time passes every 3 seconds of real time.
DayTime=13.0 means that at hour 13 (i.e. 1:00pm) you will be at 100% day factor. Prior to that, you will have something like 96% day, 4% sunrise. Setting "NightTime" to "10.3-ish" will make 10:20am or so 100% night factor. This is probably why you have night conditions at 10:30am.
You can't simply pull a TOD config from one preset and put it into another without expecting some strange behavior.

[ENBSeries]

Oyama
Here are shots of that optimization, all with SizeScale=1, SourceTexturesScale=1, SamplingRange=1, if not mentioned in here or in GUI, then quality is "high". Timing in profiler are valid for gf650ti card at 1024*768 resolution (for gf670 results will be about the same at 1920*1080).

Uncompressed images of new ssao without filter (ie no blurring of noise) and current puiblic ssao with blurring filter:
http://i.imgur.com/BjGP110.png
http://i.imgur.com/4Dd4Nor.png
As you can see, blur greatly reduce details, especially noticable with SizeScale<1, computing ssao with per pixel normals have a lot of good sides, but they are negotiated by blurring. Uh, blurring type is 2, so it's sharp enough.

These shots are both without ssao filter applied, current and new code:



These are new ssao without filter, various quality set and you may see performance impact in profiler:
http://i.imgur.com/vuLMGDv.jpg
http://i.imgur.com/D6skJpz.jpg
http://i.imgur.com/HSfh71d.jpg
http://i.imgur.com/xCRXAg9.jpg

So, what? You know why i don't want to implement it, but is it important enough or keep existing and move to absolutely different computations? And again, not decided yet, i'm narciss in this.

[Oyama]

Boris



SamplingRange=1 ...... ?!


Seriously again ? !!

Those stats....

So what ?..... Well, what I told you in PM is still valid
You rule here... your decision.
I'm all for quality and you know it.
Anyway, I think that working with Scale <1 doesn't mean a thing if users want accurate effect.
You shouldn't clamp above 1.5, 1.0, but under it ! And unclamp above..... Just kidding...
Those comparison shots are incredible about noise... and statistics

Do I think it's important ?... I think all can be.
Well, as I said many times, your actual algorythm is priceless. I achieved a balance with it, wich I find fantastic.
But, knowing you, if you feel it's time to move to GI (I'm curious about what computation method you'll use with it), and if you know it'll be better by any means, do what has to be done !

I don't *know* a single thing about your side of code, all I can do is feel and appreciate, and twist it to no end
But hell, yes, I would test the new code in my game...

Anyway, my opinion ??? : keep it !! As an alternative to anything you'll come up with, may it be GI or something else, but yeah, keep it !

[ENBSeries]

Well, probably i'm just tired from entire night coding, but now don't want to make final changes like clone current ssao and toggle it depending from presets, write another filter for low range blurring, etc. I'm always like that, losing interest after thing is work, but not made completely. Better try GI code to compare results.

[Oyama]

Sure you're tired.
Well, if you want me to test anything and trust me enough for this...
Of course, adding any toggle between different SSAO methods, with the more parameters added for total accuracy would be really something...
Ambient Occlusion and Ind Lighting have the more significant impact on this game. (AO algorythms are important on any game anyway, some dev have total ignorance and disregarding about this, most often baking in some fake method, and here they go... )
These are obligatory effect, the most important to my eyes before anything else.

Don't lose interest !

Oh, already ready to try GI ?.... This, I wanna see....
I wiki-ed several times yesterday, and it looks really interesting.

Semi-offtop :
EDIT : all that being said, call me crazy, but I find a biiiiit noise remaining in SSAO can be very interesting, rendering wise. I came to the conclusion it sort of " dries " the image, allowing for razor-sharp details on some parts. And, while I'm off-topic, I'd die to try a non-clamped-scales version of the algorythm, just to experiment with it. I have some GPUs to kill

[coffecup]

Oyama
Why do you close your K ENB page on Nexus ?
If you dont have the time to support it, you dont have to.
I know you are retired, but there is no reason to lock the page.

Sorry about offtopic, Boris.

[Oyama]

Oyama
Why do you close your K ENB page on Nexus ?
If you dont have the time to support it, you dont have to.
I know you are retired, but there is no reason to lock the page.

Sorry about offtopic, Boris.

Off-Top
Too many reasons that I don't have to talk about, especially here.
And I see no reasons to unlock it.
My interest in Boris's work is far beyond the fake popularity of Nexus and I have other ideas I didn't apply yet.
Time. I didn't retire from ENB shaders coding or from tweaking, I have too much passion for that. I retired from Nexus.
On a side note, try uploading an ENB setup with twisted shaders on Nexus, and giving no support.... Enjoy
Trying anything new on Nexus ends up, most of time in "why did you that ?", "why don't you make a realistic preset ?", "why the fuck my game won't start ?", and so....
End off-top