FSWC's back in progress!

[pwheeler]

OK guys,

here's another new reflection tweak for you to try!

in the water20.fxh change the line:

return lerp( fresnMin, fresnMax, frdot * frdot)

(This is near the top just under the define fresnMax) To this:

return lerp( fresnMin, fresnMax, frdot * frdot * frdot * frdot * frdot)

This has the effect of changing the interpolation curve between fresnMin and fresnMax. The more frdot's you multiply by the more pronounced the curve.

This shifts the high intensity reflections more towards the horizon:

BEFORE


AFTER


I think it improves the reflections quite a bit. However, on my system it seems to come with a very slight performance hit. Can you guys confirm this?

Any mathematicians out there think of a way this can be done with out all the extra code? The value of frdot is between 0 and 1 and is directly related to eye angle. the lerp command returns x + s(y - x) which is basically an interpolation between x and y. In this case x = fresnMin, y = fresnMax and s = (frdot * frdot ...). If you square frdot the interpolation become a curve. If you cube frdot the curve is more pronounced etc. etc.

Anyway, its a bit more progress!  :)

By the way, has anyone noticed that only the cumulus cloud layer is reflected... If we could find where this is decided we could alter the cloud reflections completely.

I think one of the problems is the perfection of the reflections. If we could find a way to distort the reflections more I'm sure it would look fantastic!

Cheers,

Paul  ;)

[pwheeler]

OK here's some more pics with the new interpolation mod.

This time with an extra two frdot's (seven in total!!!)
fresnMax = 0.4
fresnMin = 0
colourbase multiple = 1.8

What do you think...?

[pwheeler]

Hi guys,

just a couple more shots with this new interpolation mod. Same settings as previous. I think it looks great! It seems to have distorted the reflections a bit so they are not so pure.

[Ashton Lawson]

DUDE!

[J.]

return lerp( fresnMin, fresnMax, frdot * frdot * frdot * frdot * frdot)

i was using that line, but it just made the water got a sort of realy pale blue but with no reflective properties.

im running SP2

[J.]

never mind sorted it out now, missed the ";" of the end

[NickN]

Sorry guys I have been in transit the last day or so and just getting settled in. It will be a few days before I can get back to a normal schedule

@ Paul, the math is not going to help in that case, as least as far as i can see. You are adding operations to the passes and therefore any formula that is going to produce the same change based on the target code line will create the same result, however, its looking to me as if what are doing is increasing the opacity of the ENV or the reflection.

If that is the case it may still be attacked from the direction of texture instead of (or as well as) code

Unfortunately I am tied up with GEX right now and I just moved to my winter home in South Florida for the rest of the season so I do not have time to go over this right now but will look at it again and monitor the progress none the less.

[pwheeler]

Hi Nick,

Hope you're having a fun time in your winter home!! Wish I had one, preferable in the carribean!

Anyway, the new added code doesn't alter the opacity. What it does is make the interpolation curve between fresnMax and fresnMin more pronounced. This has the effect of pushing the transition between these two reflection values further towards the horizon. The opacity of the reflections is still determined by these two values (I lowered fresnMax to 0.4 hense the slightly duller reflections in the screen shots).

It also has the added benefit of making the wave animation effect the reflections much more, thus making them more distorted. This creates a much more believable reflection over the sea.

By the way, I played with the old tweak of mipbias to see if I could reduce the shimmering effect in the reflections. putting mipbias=4 in the fsx.cfg seemed to reduce it a bit. I'd also imagine that a higher res, anisotropic filtering and antialiasing would also reduce the shimmer if your card can take it.

Paul

[J.]

i spent a while playing around with stuff today and got some good and weird results

if anyone wants me to test some values for them im more than happy to help, my time free's up alot after Wednesday, so as i said more than happy to test a few things

good work chaps

[NickN]

Hi Nick,

Hope you're having a fun time in your winter home!! Wish I had one, preferable in the carribean!

Anyway, the new added code doesn't alter the opacity. What it does is make the interpolation curve between fresnMax and fresnMin more pronounced. This has the effect of pushing the transition between these two reflection values further towards the horizon. The opacity of the reflections is still determined by these two values (I lowered fresnMax to 0.4 hense the slightly duller reflections in the screen shots).

It also has the added benefit of making the wave animation effect the reflections much more, thus making them more distorted. This creates a much more believable reflection over the sea.

By the way, I played with the old tweak of mipbias to see if I could reduce the shimmering effect in the reflections. putting mipbias=4 in the fsx.cfg seemed to reduce it a bit. I'd also imagine that a higher res, anisotropic filtering and antialiasing would also reduce the shimmer if your card can take it.

Paul

72 degrees and light rain, I'll take it, especially when it will be back to 79 and sunshine in a day or so

My comment about reducing the opacity was rushed on my part and not well stated however I am in tune with what you are working with there. If there is a way to make the actual reflection texture rework that surface in conjunction with that code, as I briefly touched on here: http://www.simviation.com/cgi-bin/yabb2 ... 4933/34#32

we can probably place more control in the tweak.

Have you tried Nhancer and setting the system up as I have noted here: http://www.simviation.com/cgi-bin/yabb2 ... 53/165#172

You may find most if not all your shimmers are gone.

[Ashton Lawson]

Nick, regarding the nHancer thing, I think I can do that too, with FSWC.

nHancer only edits an xml file under WINDOWS or system32.

[NickN]

[quote]Nick, regarding the nHancer thing, I think I can do that too, with FSWC.

nHancer only edits an xml file under WINDOWS or system32.

[Ashton Lawson]

k then

EDIT:

I've just installed the FSX Accel SDK.  it even has support for 3ds Max 9, convenient for me.

anyway, i'm gonna look through the variables list and stuff, see what i can find..

[NickN]

Ok guys, here is your variable... try:

#define fresnMin 0.05
#define fresnMax 0.39

half fresnel2( half3 vEyeDir, half3 vSurfaceNorm )
{

[Ashton Lawson]

man, that's good.

still, it would seem it would no longer be changing the curve, more like lowering it.  I suppose we could implement both, to give more control.