UPLOADED: Ankor's Shaders Update 22 October 2019!

Yes. Didn't realize the readme wasn't in there, just grabbed it out of my folder of vital add-ons never to lose. Thanks for posting that James.
 
gecko said:
Joost, have you installed the 4gb patch to your cfs3.exe? It probably isn't it, but it's worth a shot. I was trying to think what might be different by default with WOTR/WOFF where this problem never surfaced over there, and this was what came to mind.
View attachment 72090
Click to expand...

I think I had that installed, but tried it anyway, to no avail - unfortunately. The file you sent me through email didn't work either :sorrow:
But on my laptop, even with just 17 FPS, both the shaders and the Spit work. So it's gotta be something with my Win7 x64 rig...
 
And I've already made a test version. It was easier than I expected.

Download here: https://1drv.ms/u/s!AvmtwF9vPtwVgcM5znXuECtvkaN-jA?e=a4nTW6

Only d3d8.dll in the archive, just use with the new shaders.

If it crashes, it will save "d3d8.dmp" in CFS3 folder. This file along with logs should help me investigate the cause.
I regret not implementing this feature earlier. Could have saved me some headache.
 
That is fast!

Here are the results: View attachment Shaders Log & Dump.zip

I started CFS3 ETO with TiR activated but without the Spit. I also did a memory check (which reported no faults) but that shouldn't be it, as ETO runs fine without the shaders.
Thanks very much for looking into this, Ankor :encouragement:.
 
Frosty, can you try this version: https://1drv.ms/u/s!AvmtwF9vPtwVgcM-0FafflnYMg3oDA?e=deCANi

The dump shows that an error is inside Direct3D code, so either it is a bug in a gfx driver or my code does something wrong that corrupts the runtime and causes a crash.
Unfortunately I'm on Win10 and DirectX 9 debugging is not supported since Win 8.1 so I can only hope that I'm not doing any stupid mistakes.

Anyway I made some (rather minor) changes that may fix the problem, but I need you to test this.

I've also found and fixed a couple of unrelated bugs that weren't noticed before: water animation is not smooth, and control options may show duplicate entries for the same command if it is added from multiple ini files.
 
Thanks Andrey! Can't confirm if it fixed Frosty's problem obviously, since I never had it, but this version is working well for me, and the smoother animation of the water is nice!
 
AnKor said:
... Anyway I made some (rather minor) changes that may fix the problem, but I need you to test this. ...
Click to expand...

Thanks!! I'll give it a go (but I am not sure I can find the time today).
If you say 'Direct3D code' and/or gfx driver', would reinstalling DirectX or updating/changing my graphics card drivers help?
 
Reinstalling DirectX is unlikely to change anything, but a different graphics card driver may help.

And it just occurred to me that my wife's old notebook still has Windows 7! :)
It is 11" i3 from 2011 and can barely run my shaders, but at least I'll be able to verify that they don't do anything invalid.
 
AnKor said:
Here is a new version (again): https://1drv.ms/u/s!AvmtwF9vPtwVgcNAHdHQnWR1TVz6uw?e=WmsYlK

This time I've used Win7 to debug and immediately found a bug with in-cockpit reflections code which is very likely the culprit.

Frosty, please ignore the yesterday's (Nov 07) build and try this one :)
Click to expand...

Yes, yes and triple YESS!! That seems to have done the trick!!
I now have the shaders working and the new clickable cockpit-Spit fired up! That is just great work! :encouragement: :applouse: :loyal:.

I do have one question about the setting up of joystick buttons with the clickable cockpit and I don't know if this one should be adressed to you, Ankor, or to Daniel. There is the following text in the "controlassigments.ini" -file:

"Joystick buttons can be assigned as well. The naming convention is XJoyY, where Y is the Joystick number and X is the button number. If you have a single controller plugged in, Y will always be 1. If you have multiple controllers plugged in you may need to try several joystick numbers to find the correct joystick number for your desired button. For example, if you want to assign the Brake Control to button 6 on your joystick, but you also have a set of rudder pedals and a separate throttle quadrant, start by trying 1Joy6, and if that doesn't work, 2Joy6, and so on until the button works"

I read the quote as "The naming convention is XJoyY. X is a button on controller Y. If you have two or more controllers and you want button 6 assigned to the Brake function, you have to try switching controller numbers to find the correct controller while keeping the button number the same (= 6)". I could be misunderstanding things, but shouldn't I try 6Joy1, 6Joy2 and 6Joy3 if I wanted to find button 6 on one of 3 different controllers?
BTW, is the button number the same as the DirectX button number? (Yes, I can't find the Brake function on my controllers... :biggrin-new:)
 
Frosty, glad you are enjoying it :)
Regarding joysticks, it is indeed a question to Daniel, but for me it also looks like it should be 6Joy1, 6Joy2, 6Joy3 as you say.

gecko, since the latest dll works well, I think you can just add it to shaders package and reupload it.
Changes are:
- fixed possible Win7 crash;
- water animation is smooth (as it was in 2018 version);
- Numpad0Switch=1 can be used again (if you really need it);
- dll writes a dump file if the game crashes.
 
You don't need to do anything to the ini file. That file is used only if you don't have the shaders installed. It is unnecessary with the shaders installed because all key configuration is handled through the normal in game key configuration menus.

I'll get the updated shaders together and reupload the complete package soon.
 
My apologies, but a lot of this is over my head, so I have to ask if the issues being addressed in this thread have been ironed out and if a revised package is going to be uploaded? The 'fake driver' thing has always been key for me.
 
Yes, it appears everything has been corrected. I hope to have time to put up an official revision tonight or tomorrow night.
 
Need some help installing these new shaders please.

Installed 2019 shaders using gecko's Oct 29 link, Ankor's Nov 8 d3d8.dll and Major Magee's d3d8.ini into win 7. Everything seems to work ok except the dynamic gunsight. This shader package was installed over the Nov 2017 version using jsgme. I did not remove the previous version which was not added with jsgme. If I remove this version with jsgme the dynamic gunsight works again. I tried to figure out what section 6 in the developer notes said about this but I don't follow what instruction 3 is saying.

I want to get this working so I can add the new spitfires with all their new features.
 
Now have the dynamic gunsights woriking. I had deleted a lot of the effects.dds files in the Shaders30 texturemagic.ini file when I installed the new shaders package. I had installed from the unzipped file at the upper levels so it overwrote the complete folders and files. I had to spend a lot of time combining the old and new, making sure I got everything, without getting duplicate files and tried to keep it in reasonable order. At least I had used jsgme so was easy to put back.
 
I've been trying out the latest 4.4.2.700 version of Reshade this evening with the new shaders, and there doesn't seem to be any inherent compatibility issues. A few of the Reshade Effects cause the game to crash, but most work just fine. They all have varying degrees of negative impact on the FPS, but the Reshade interface makes it really simple to turn them on and off in game. The install is really simple (run the downloaded exe, point it at the cfs3.exe, select the Direct3D9 button, and let it download all the available shader effect files, you can ignore the message about needing the additional wrapper file when using AnKor's dll) and at this point I don't see a need to go back to using SweetFX again.

https://reshade.me/

P.S. If the game crashes you'll have to hand edit the Reshade.ini to prevent it from reloading the offending effect on start-up the next time, and immediately crashing out again.
 
Last edited:
I just noticed something I don't usually pay any attention to. With these new shaders, independent of the setting selected in cfs3config.exe, the game reports that it's running in 16 bit color. If I try to change it in game to 32 bit I get an error message about having insufficient graphics memory, and it crashes. cfs3config.exe doesn't even offer the 16 bit option (32 bit only), yet that's what shows up in game after you run it. The insufficient memory crash appears to be completely independent of the resolution being selected.

With the previous shaders I can run the game either way (and both are available in cfs3config.exe), though I can't see that it makes any visible difference to the graphics that are being displayed (screen capture bmp files are 24 bit in both cases).

Is this some sort of reporting error, or is the game actually running with reduced colors with these new shaders?

==========
Microsoft Combat Flight Simulator 3
build 030203.01 run at Mon Nov 25 17:27:35 2019
==========
CFS3: EnableAsserts: true
CFS3: ExitOnAssert: false
CFS3: SingleAssert: false
CFS3 ERROR [<graphics> cdirect3d(430)]: There is not enough free video memory to run. To correct this problem, reduce your desktop resolution or change the bit depth from 32 to 16.
Click to expand...
 
You must log in or register to reply here.
Back
Top