Attention all skinners!

[AnKor]

The problem is that I don't keep such list. I know it is a bad practice, but the way I work on this mod is more like "oh, a bug, let's fix it... oh, I don't like how that part looks, let's adjust it... etc".

Still, from what I recall in no particular order.

1. Corrected dynamic reticle texture orientation.
2. Fixed weird graphics glitches and odd look of textures with alpha when dynamic reticle is visible.
3. Fixed frequent crashes after Alt-TAB when extended textures were used.
4. Fixed bright 1-pixel line appearing on the ground at certain distance and viewing angles.
5. Fixed case-sensitivity for TextureMagic.ini
6. Cloud shadows are reset when starting new missions. Previously there were some leftovers (wasn't an issue for WOFF as there is always just one mission per session).
7. The log now reports cfs3.exe version - may be useful for troubleshooting.
8. Dynamic reticle zoom switchable.
9. All objects which use textures from "buildings" folder now cast real shadows as part of scenery instead of stock cfs3 ones.
10. Improved ground self-shadowing code (shadows from hills) to avoid artifacts. Though there seems to be some issues.
11. Big overhaul of aircraft lighting - an attempt at physically based shading. Better use of fresnel effect, specular texture now control sharpness and brightness of highlights. If no specular map is available (like in WOFF) the highlight details are inferred from aircraft texture (similar to how bump mapping works). I will explain details later, first I want to know whether you like the new look
12. Cockpit and other light brightness is adjusted according to ambient lighting. I.e. lights are less bright at mid day. Cockpit light again uses red as its *specular* color (stock cfs3 uses pure white).
13. Brighter ambient light at evening/morning -- otherwise terrain was too dark. Some blue tint may be there as fog color is now mixed into ambient color.
14. Some attempts to improve game stability -- my code now tries to recover from some errors instead of crashing right away

Known Issues.
1. Once again highly reflective metallic and glass surfaces may be off. Less so for metallic, but reflection effect is very weak on glass. I had it working at one moment but with new lighting model I have to redo this effect.
2. Dynamic reticle appearing behind some cockpit objects (P47).
3. Shadows from transparent objects - I haven't looked into it in detail, but I believe WOFF relies on this oddity for some models by placing normally invisible pilot's outline into the cockpit.
4. Normal maps doesn't seem to work properly, may cause odd visuals like inverted lighting.
5. Gloss maps are not supported (I haven't explained what they were for anyway ) -- with new lighting model specular maps are used for glossiness.
Also this specific version uses experimental terrain lighting - it will be reverted to previous one, so ignore for now.
6. Low sun lighting and sky/fog brightness still needs some tuning.

But as always there might be something else I forgot
I'm not asking to test it thoroughly or look for specific bugs, any kind of feedback is appreciated.

[MajorMagee]

The variable reticle size works wonderfully. It's going to be a fair amount of work to calibrate it, but what a realistic addition to the game it will be once I have it done.

I have not noticed any of the graphics glitches that were cropping up in the game before. What I did see was that a few of the reflective aircraft (P-38, P-51B) did not display correctly in the Selection Screen at the airfield, but look perfect in game. It shows up as bands or small areas of the skin that are highly reflective, while the remainder of the skin looks normal. I've looked at the reflection map dds and the display settings in the m3d file, and nothing appears amiss. My guess is that it's something like a surface normal that's flipped in the original 3D model. It may be that you have it patched for the in game display, but that fix in your code is not being applied to the 3D UI world.

I did notice that the new lighting effects make the color range of the skins more dynamic which in some cases is a good thing. In direct light the light colors get lighter and the dark colors get darker. I'm not sure that's actually the way it should work. I really noticed it on a Navy Blue aircraft that was an overall weathered Blue Grey, and it turned a Darker Blue when it was in the sun rather than a brighter Blue Grey. I'm sure that there are some parameters in one of the .fx files that could be adjusted to keep the dark colors from going the wrong direction when they sun hits them

[AnKor]

Hi!

I'm thinking about adding a "developer mode" for gunsight placement. You will be able to move and scale the reticle sprite in-game and see exact coordinates in real time. This should make it easier to set up for various aircrafts.

I see the problem with P-38L, will investigate. However it is very odd as there should be no difference between selection screen and game rendering, so I can't even guess what is wrong until I look into it.
You are right though that on some models (including WOFF ones) surface normals are flipped or otherwise wrong. It wasn't very noticeable with the stock renderer but once I started adding reflections such issues started to appear in various places. One notable case is stock Spitfire IXc/e cockpit: there are 3 "thingies" (half cylinder and a box) in upper front of the canopy frame - you can always see them when looking forward with normal zoom. And the rightmost one has flipped normal on the "box" part - the lighting is inverted - unless it is in shadow you can see it becomes bright when other two are dark and vice versa

As for color issues - I think I understand what you mean, but can you post a couple of screenshots comparing my new and the stock renderer. There might be various reasons for such color change so I need to be sure.

[gosd]

MajorMagee;
When using the [ReticleDynamic] ; scale|SDF do I have to put in the scale (ie .dds=xx) for every reticle I will use or did you just gave that as info?

[MajorMagee]

Each one get's it's own definition, so they can be tuned to matched each specific aircraft's gunsight.

[gecko]

I've had only a quick look at it so far, but I like what I see.

Interestingly, the reflections come in strong on the glass of the beta Me 163 project I'm working on. It has a mostly transparent texture applied to the glass to provide a little detail, which may be the reason, as most other aircraft do not have textured glass and are as you described, very dark and only a faint reflection. Anyways, I know that's a known issue but thought you'd like the extra info.

I like the way the fresnel looks, it's really good. I can't be sure yet, but specular textures seem to have some of the detail washed out compared to previous versions, but Ill need to look closer, that may not be the case.

The only negative so far is that I seem to get slight stuttering and about 10 less FPS compared the same settings on the last version. I typically got 29-30 fps before, now I get about 19.

I hope to get some more time with it this weekend. Thanks!

Daniel

[MajorMagee]

Here are a few examples of the darkening effect. Each image on the left is with AnKor's shader toggled off, and the one on the right is with it turned on.







Most aircraft don't experience this color shift, and actually look much better than with the stock shading, so it's really just a minor curiosity.

[gosd]

I find the shadows of buildings and trees to be a bit exagerated.

[AnKor]

Interestingly, the reflections come in strong on the glass of the beta Me 163 project I'm working on. It has a mostly transparent texture applied to the glass to provide a little detail, which may be the reason, as most other aircraft do not have textured glass and are as you described, very dark and only a faint reflection. Anyways, I know that's a known issue but thought you'd like the extra info.
Ok. Hard to tell now what is the difference.

I like the way the fresnel looks, it's really good. I can't be sure yet, but specular textures seem to have some of the detail washed out compared to previous versions, but Ill need to look closer, that may not be the case.
Specular texture now work differently - instead of just telling how bright the shine is, they control the sharpness of the specular directly, and brightness indirectly because with new shaders these are interconnected.
That's why they may appear different to what you are used to. However, I'm still going to adjust that a bit so it may improve.

The only negative so far is that I seem to get slight stuttering and about 10 less FPS compared the same settings on the last version. I typically got 29-30 fps before, now I get about 19.
I hope to address that. However as I recall you have GTS 250 which is rather old by now.

Here are a few examples of the darkening effect. Each image on the left is with AnKor's shader toggled off, and the one on the right is with it turned on.

MajorMagee, oh, that's very noticeable.
My only idea so far is that those models might use non-white (like blue or brown) specular color with low glossiness and my new shaders assume that the specular is always white and also increase overall glossiness. Realistically it is always the case with non-metal surfaces, only metals have colored reflections - gold tints everything as yellow, etc. And I guess an ordinary paint (without metallic effect as used on cars) should have white reflections also.
Can I get any of these models to check?

Another thing I noticed is that models which use both _r and _s textures at once don't work properly now (and maybe never worked before). If I recall correctly such arrangement is not standard but still supported by cfs3 so I will look into it.

I find the shadows of buildings and trees to be a bit exagerated.
gosd, too dark or too long?
I mentioned that these new shaders modify the ambient value at evening/morning to reduce shadow contrast, but at the same time slightly decrease the ambient at midday to make shadows darker.
As for the length - realisitcally shadows should become blurrier with the distance, but right now they do not, that's why they might look weird sometimes. I can implement it, but it will eat quite a few FPS.

[gosd]

Ankor
I find the shadows of buildings and trees to be a bit exagerated.

The shadows are too long. And it does not change depending of time of day.
Also I noticed that when the effects are on the pencil of ray from the search lights are hardly visible on my setup.

[MajorMagee]

Each of the examples above were actually a problem with the way they looked with AnKor's Shaders switched off, and not when they were activated.

The F4U-4 and G50 had a Glossiness setting of 00 00 00 40 instead of my standard 00 00 00 42 in their m3d files. Once I changed it to 42 everything looked correct again.

I found a color bias in the specular factor (slightly redder than neutral at 30 30 42 so it made the dark green look brown) in the Ar 234's m3d file that explained the effect. Once I corrected it the issue cleared up for that one as well.

[MajorMagee]

I know you're just experimenting with the ground shading methods, but the latest version seems to make the Anisotropic Filtering less effective. Even at 16X there's a fair amount of crawling of the tree shadow renderings as you move around the scene. Perhaps it's just that there's a more distinct edge to the shadows?

[MajorMagee]

I've been working on calibrating the multiple magnifications for each gunsight, and have finished about half of them. Most of them I'm setting up to display fighter, medium bomber and heavy bomber sizes all set to coincide with 400 m. It will be easy enough for anyone to fine tune these to suit their tastes, but I did want to give a reasonable starting point for those who don't want to fuss with adjusting it. It's interesting that as I'm trying to get the range of the fighters by coming up on their six until I'm 400 m away they inevitably break when I'm about 450-425 m behind them.

Alternatively, some may want to pick out just one EA type and the presets could be for that particular target at different ranges.

I'm finding this new functionality to be a particularly nice addition for enhancing the immersion factor. Even for those who do not have TrackIR to take advantage of the dynamic reticle movement they will still be able to benefit from being able to adjust the size during the mission to match their different targets.

[Daiwilletti]

Sounds brilliant. Am looking forward to seeing these reticules.

thanks!

[gosd]

Ankor
I find the shadows of buildings and trees to be a bit exagerated.

The shadows are too long. And it does not change depending of time of day.
Also I noticed that when the effects are on the pencil of ray from the search lights are hardly visible on my setup.

How can I correct these issues?

[MajorMagee]

Since AnKor made his Beta available (post #121 on page 5) earlier this month I decided to go ahead and package up what I already have completed for others to start using.

The Dynamic Reticle package will be available here once it's published:

http://www.sim-outhouse.com/sohforums/local_links.php?catid=13&linkid=19730

If you have any trouble with the instructions (it only requires some text editing) just ask for help, and I'll try to walk you through it.

As I found during the development of this collection, you can implement this conversion a little at a time once you have the first couple of steps in place, so don't be put off by thinking that you have to do hundreds of planes all at once to use it.

[HouseHobbit]

Since AnKor made his Beta available (post #121 on page 5) earlier this month I decided to go ahead and package up what I already have completed for others to start using.

The Dynamic Reticle package will be available here once it's published:

http://www.sim-outhouse.com/sohforums/local_links.php?catid=13&linkid=19730

If you have any trouble with the instructions (it only requires some text editing) just ask for help, and I'll try to walk you through it.

As I found during the development of this collection, you can implement this conversion a little at a time once you have the first couple of steps in place, so don't be put off by thinking that you have to do hundreds of planes all at once to use it.

Thanks Major, these are Most welcomed.

[MajorMagee]

I just mapped the DPC Mig-15 to use:

<Effect Type="Track" EffectName="RU_USSR_LA5_gunsight" PosX="-0.025" PosZ="0.95" PosY="0.930" Pitch="90" MinVel="-999999"


MaxVel="999999"/>

[gecko]

Very nice Major! I'm curious, what was your process for setting the reticle sizes. Finding historical sizes has been difficult, and then even harder to translate to CFS3 usage.

[ndicki]

Quick pop baqck to say hello and see what's going on!

I've just installed the latest beta shaders pack and modded exe into MAW, and am extremely impressed! Very, very good. My FPS are at 50, which is the upper limit I set, with full 5s on all sliders.

Only one niggle that I can see - spinner rotation on the Bf109e is a bit odd. You can see this on any of the aircraft with two colour spinners. Otherwise the new viewpoint is brilliant, and the new shaders a whole new experience!

[MajorMagee]

Typically I selected several common aircraft that represented the different size targets (e.g. Bf109, Ju88, Fw200). I would fly a QC mission and come up from behind until I was 400 m away to check the reticle size. I'd then change the magnification multiplier, and do it again until I had a good match. A few historical documents suggest that the distance should have been 300m, but I found that with the CFS3 models it often led to the bomber wingspans being well outside the width of the gunsight reflector glass. My philosophy generally ended up being to give an indication that you had gotten close enough to be within the maximum effective range when the wingspan just covered the full circle. The ratio between the types really doesn't change, so if you wanted to move them in closer to the gun convergence distance (250-300m) you would just proportionally increase the multipliers I used.

[MajorMagee]

I picked up a copy of "Dogfight: The Greatest Air Duels Of World War II" a couple of months ago, and it was fairly helpful with this issue.

from page 255

You must fire your weapons only at the ranged for which they were designed. Normally, you should not fire at ranged greater than 400m, as beyond this distance the trajectory quickly falls away. This also applies to larger weapons such as the Mk 108. It is a common fallacy to believe that it must be possible to fire larger-caliber weapons at greater ranges, and that the aiming of these need not be as precise as for small-caliber weapons. The opposite is true! You have little ammunition! Use it sparingly. With the Mk 108 you only have 60 rounds, so reduce range, aim carefully, fire accurately.

Commence firing at a maximum range of 400m. Experience has shown that combat ranges are being greatly underestimated. The distance given in combat reports are almost never correct. When, for example, a combat range of 50-100m is reported, it actually varies between 200m and 400m, as established by analysis of combat films. Frequently, range estimation errors are even greater. In attacks against bombers, many fighters open fire at 2,500-3,000m. This is a senseless waste of ammunition!

A four-engined bomber has a wingspan of about 30M. If it appears a large as the deflection circle it is at a 300m range. (I set mine up for 400m so you might want to change the multipliers) For a four-engined bomber, the estimated value corresponding to a span of 30m is multiplied by 3. Therefore, a four-engined bomber with 30 m span fits in the circle one time - 1 x 3 = 300 m; fits in the circle two times - 2 x 3 = 600 m, fits in the circle three times - 3 x 3 = 900 m.

July 1944 Gunnery Instruction Document

It really is hard to hold off from firing too soon, even with the sights calibrated for 400m. As I mentioned earlier, many of the enemy aircraft have the annoying habit of breaking away as soon as you get inside 450M.

[gosd]

Ankor
I find the shadows of buildings and trees to be a bit exagerated.

The shadows are too long. And it does not change depending of time of day.
.

[gosd]

Forget the first photo, the one with the Swordfish illustrates what I am talking about,.

[italflyer]

OK, guys, what am I doing wrong?

Tried to add the dynamic reticle to the Italian Wings MC200 (IW_MC200_356-13) by copying the Line from the G50 into its .dpd and changing it accordingly. The line nos read like this:

<Effect Type="Track" EffectName="IT_CAI_Tipo_C_gunsight" PosX="0" PosZ="-1.2" PosY="0.94" Pitch="90" MinVel="-999999" MaxVel="999999"/>

The position of the dynamic gunsight seems quiet correct, but the old static is still there. Here's a screenshot of what I get:



Any Ideas?