WIP versions

[JangoJungle]

Hi, I'm new here . I've been testing dgvoodoo for some time now and I have tested some old games like Nightmare Creatures and Resident Evil 2 SOURCENEXT version. Resident Evil 2 SOURCENEXT booted in the WIP37 but could not reproduce the SOURCENEXT intro video and crashed. Now on WIP43 it reproduces but only audio without video but is finally PLAYBLE.

[batterymandark]

dgvoodoo can't play certain video in fullscreen, you have to force the game to run in windowed mode, but you can run it in your desktop resolution with the correct aspect ratio

[JangoJungle]

dgvoodoo can't play certain video in fullscreen, you have to force the game to run in windowed mode, but you can run it in your desktop resolution with the correct aspect ratio

I have tested a mix of settings today, like the windowed mode one, several times but with no luck.

[vedmysh]

A new WIP (would be fine to be the last before the next version):
Well, of course it's on my list.
However those game series have one problems that seem to be unresolvabe to me for the time being: additional windows on top of the rendering window.

Nice to hear that.

Yeah, that wierd thing with two windows DxWnd somehow patched/managed this issue. But DxWnd have terrible fps dropdowns in "graphics emultation" part. I tried to use DxWnd in collaboration with dgVoodoo without any luck

[Myloch]

Webdiver runs correctly with ATI card, it refuses to run with nVidia card (with a pop-up window mentioning a numeric error).

It now works with latest wip (43). Thanks.

[Dege]

Thanks for the feedback to everybody, here too!!

@Batterymandark: thanks for your videos, I really appreciate it. I plod through the installing of TOCA Racing Driver on Win10 and finally managed to run it. I fixed the depth-issue and another one I found (crash when entering a subsequent race).
The only flaw now is the game crashes at exit (incompatibility coming from modern DX).

Hi, I'm new here . I've been testing dgvoodoo for some time now and I have tested some old games like Nightmare Creatures and Resident Evil 2 SOURCENEXT version. Resident Evil 2 SOURCENEXT booted in the WIP37 but could not reproduce the SOURCENEXT intro video and crashed. Now on WIP43 it reproduces but only audio without video but is finally PLAYBLE.

Yes, video playback is a known issue with dgVoodoo. I've been planning for a long time to address it.

[vedmysh]

Dege, i've done some quick static analysis on Jane's F18 executable. Not sure, will this help (hope it is) or not, anyway, here we go:

The game consist of two parts (and looks like of two windows):

"wrapper" (gui, mission planning, etc) and "sim" - simulator and rendering part

All VA is from "safedisk-free" executable 1.01f, image base at 0x400000

"sim" window is main window created at .text:00542070 (called from WinMain)

not sure about "wrapper" window, but looks like it created in wrapper.dll - main gui module with heavy mfc usage

window procedure located at .text:00541760 and is relativly small.

on window message 0x500 (sent by PostMessageA call) executed transition from "wrapper" to "sim"
on window message 0x501 (sent by PostMessageA call) executed transition from "sim" to "wrapper"

from "wrapper" to "sim" transition procedure located at .text:00541C70 and calls:

ShowWindow("sim window" handle, SW_SHOW)
UpdateWindow("sim window" handle);
free wrapper stuff
start sim

from "sim" to "wrapper" transition procedure located at .text:00541CF0 and calls:

ShowWindow("sim window" handle, SW_SHOW)
UpdateWindow("sim window" handle);
...
init wrapper stuff
ShowWindow("sim window" handle, SW_HIDE);

Also on window message 0x503 wrapper dlls are initialized. This happends at startup in WinMain:

.text:00540924 push 0
.text:00540926 mov ecx, [ebp+wParam]
.text:00540929 push ecx
.text:0054092A push 503h
.text:0054092F mov edx, main_gui_window
.text:00540935 push edx
.text:00540936 call ds:PostMessageA

On window message 0x504 wrappers stuff uninitialized.

More interesting windows managment in window procedure:

.text:0054182C mov wm_activateapp_flag, ecx
.text:00541832 cmp wm_activateapp_flag, 0
.text:00541839 jnz short loc_54185A
.text:0054183B mov edx, main_gui_window
.text:00541841 push edx ; hWnd
.text:00541842 call ds:IsWindowVisible
.text:00541848 test eax, eax
.text:0054184A jz short loc_54185A
.text:0054184C push 6 ; nCmdShow
.text:0054184E mov eax, main_gui_window
.text:00541853 push eax ; hWnd
.text:00541854 call ds:ShowWindow

case WM_ACTIVATEAPP:
flag = (wParam == 1 || wParam == 2);
if (!flag)
{
if ( IsWindowVisible("sim window" handle) ) ShowWindow("sim window" handle, SW_MINIMIZE);
}

DirectDraw/Direct3D initialized from .text:005728E0 procedure, which is called from some global initialization function (.text:005730C0),
which is called from "start sim" part (.text:00571C30)

[batterymandark]

Thanks for the feedback to everybody, here too!!

@Batterymandark: thanks for your videos, I really appreciate it. I plod through the installing of TOCA Racing Driver on Win10 and finally managed to run it. I fixed the depth-issue and another one I found (crash when entering a subsequent race).
The only flaw now is the game crashes at exit (incompatibility coming from modern DX).

Hi, I'm new here . I've been testing dgvoodoo for some time now and I have tested some old games like Nightmare Creatures and Resident Evil 2 SOURCENEXT version. Resident Evil 2 SOURCENEXT booted in the WIP37 but could not reproduce the SOURCENEXT intro video and crashed. Now on WIP43 it reproduces but only audio without video but is finally PLAYBLE.

Yes, video playback is a known issue with dgVoodoo. I've been planning for a long time to address it.

That's really cool that TOCA was an interesting candidate , and yeah the most pain of installing TOCA race driver is that on windows 10 64 bit it need both [HKEY_LOCAL_MACHINE\SOFTWARE\Codemasters\Race Driver] and [HKEY_LOCAL_MACHINE\SOFTWARE\WOW6432Node\Codemasters\Race Driver]

So for me i installed toca on a virtualbox and copied files over to my main os , and "killed secure rom" , since its useless on windows 10

[KainXVIII]

About Blade of Darkness - with D3d rendered and latest wip game looks pretty good (8xMSAA), but despite that fps counter shows 60 fps, game not really feels smooth, slight stutters here and there..
On other hand with Nglide 2.0 and r3dfx v1-2 renderer game is smooth, but has no mipmapping support, no msaa, colors looks a little washed out and flames sometimes lights through walls - decisions, decisions!

[VicRattlehead]

About Blade of Darkness - with D3d rendered and latest wip game looks pretty good (8xMSAA), but despite that fps counter shows 60 fps, game not really feels smooth, slight stutters here and there..
On other hand with Nglide 2.0 and r3dfx v1-2 renderer game is smooth, but has no mipmapping support, no msaa, colors looks a little washed out and flames sometimes lights through walls - decisions, decisions!

I would guess that it's because of the way the light sources are rendered. I get severe framerate drops with the D3D renderer when any light source is on screen. Maybe your PC is powerful enough to handle them without dropping below 60 FPS but the performance hit still manifests itself. I switched the game's renderer to "D3d i740 Windows98" which changes the rendering a bit (no lens flare is the only difference I noticed) and the game plays smoothly even on my weak PC.

[KainXVIII]

About Blade of Darkness - with D3d rendered and latest wip game looks pretty good (8xMSAA), but despite that fps counter shows 60 fps, game not really feels smooth, slight stutters here and there..
On other hand with Nglide 2.0 and r3dfx v1-2 renderer game is smooth, but has no mipmapping support, no msaa, colors looks a little washed out and flames sometimes lights through walls - decisions, decisions!

I would guess that it's because of the way the light sources are rendered. I get severe framerate drops with the D3D renderer when any light source is on screen. Maybe your PC is powerful enough to handle them without dropping below 60 FPS but the performance hit still manifests itself. I switched the game's renderer to "D3d i740 Windows98" which changes the rendering a bit (no lens flare is the only difference I noticed) and the game plays smoothly even on my weak PC.

But fps still shows 60 without drops for me (maybe vsync or fps cap is broken in this game).
Found thread with color difference screenshots https://www.gog.com/forum/blade_of_dark ... ould_i_use

[VicRattlehead]

Age of Wonders: Shadow Magic regression?

2.54 = works fine
WIP43 = black screen and audio only in fullscreen mode

Also, not related to dgVoodoo2, but the CPU usage of this game is quite heavy for some reason. Can't really enjoy playing it as it is.

EDIT: Noticing a similar problem with Warlords Battlecry 3 too. With 2.54, the game works fine, but with WIP43 I just get a black screen and this time the game doesn't even seem to boot properly because I don't hear audio and I have to kill the process with the task manager.

[Dege]

I've finished testing the new WIP with hundreds of demos, games and such, and also fixed some regression issues I encountered:

http://dege.fw.hu/temp/dgVoodooWIP44.zip
http://dege.fw.hu/temp/dgVoodooWIP44_dbg.zip

=========================
WIP44: (next final)
=========================

- D3D FPU state handling incompatibility fix (for general cases, and it fixed nVidia demos Creature, Toy Soldiers and Crystal Ball)
- DDraw surface-create and other incompatibility fix (Zero Comico, RC de Go)
- Gunmetal (Glide) on AMD fix
- D3D8: implementing ValidateVertexShader and ValidatePixelShader for Microsoft Shader Assembler (Shadow of Destiny)
- D3D8: some object handling incompatibility fix (TOCA Racing Drive)
- DirectX: fix for rendering with incompatible rendertarget/depthstencil buffers (TOCA Racing Drive)
- Some minor internal thing is optimized/refactored/fixed
- Some effort for avoiding app deadlocks in DDraw and QuartzHookLayer
- Minor internal D3D state/lighting bugs fixed (Tonko4)
- Adding 'DesktopResolution' and 'FreeMouse' as new options to GeneralExt section
- D3D colorkey bug fixed (Sponge Bob - Employee of The Month)
- Some regressive issues are fixed

If no regressive bug reported then I'll release it soon as a new 'official' version.

Age of Wonders: Shadow Magic regression?

I've fixed it in this WIP. But, this game works only properly with no forced resolution and 'unspecified' scaling mode.
The game relies on physical window size and produces black screen when it's other than the game expects.
Interestingly, the same issue as with Jagged Alliance.

[batterymandark]

I've finished testing the new WIP with hundreds of demos, games and such, and also fixed some regression issues I encountered:

http://dege.fw.hu/temp/dgVoodooWIP44.zip
http://dege.fw.hu/temp/dgVoodooWIP44_dbg.zip

[i]=========================
WIP44: (next final)
=========================

- D3D8: some object handling incompatibility fix (TOCA Racing Drive)
- DirectX: fix for rendering with incompatible rendertarget/depthstencil buffers (TOCA Racing Drive)
If no regressive bug reported then I'll release it soon as a new 'official' version.

Age of Wonders: Shadow Magic regression?

I've fixed it in this WIP. But, this game works only properly with no forced resolution and 'unspecified' scaling mode.
The game relies on physical window size and produces black screen when it's other than the game expects.
Interestingly, the same issue as with Jagged Alliance.

I can gladly say that the depthstencil buffers fix seems to doing its job, "as i guess from the name"
no more ghostly buildings and metal grids.
And the game run extremely well on a GTX1080 at 5120x2880 , 240+ fps when driving alone and 100+ fps when driving with all cars.

Also i hope its going well and that the fixes for TOCA doesn't only apply to my request, but will also fix other games in the same turn.

[Dreadmoth]

The colorkey changes might have some negative effects when forced texture filtering is used:

Machines

App driven
Forced bilinear (previous versions of dgVoodoo 2)
Forced bilinear (WIP44)

[Dege]

Also i hope its going well and that the fixes for TOCA doesn't only apply to my request, but will also fix other games in the same turn.

Handling incompatible color/depth targets had already been there in dgVoodoo (because of some other games like Zanzarah) but TOCA needed a more improved version of that.

The colorkey changes might have some negative effects when forced texture filtering is used:

Machines

App driven
Forced bilinear (previous versions of dgVoodoo 2)
Forced bilinear (WIP44)

Thanks!
Sadly however, the one present in WIP44 is the correct one.
The colorkeying bug was that dgVoodoo treated colorkey pixels as black transparent ones (like for colorkey blending) which caused black edges in other game(s) where it shouldn't be. But undoubtedly it came handy for Machines.
What if you try one of the GeForce virtual videocards? They feature 'full cut colorkeying' which means no partially colorkeyed pixels should appear anytime, but in turn, colorkeyed objects are 'thinner' and jaggier.

[Dreadmoth]

Sadly however, the one present in WIP44 is the correct one.
The colorkeying bug was that dgVoodoo treated colorkey pixels as black transparent ones (like for colorkey blending) which caused black edges in other game(s) where it shouldn't be. But undoubtedly it came handy for Machines.

Yeah, when I first tried forced filtering I half-expected to see more issues with the colour-keyed textures. Makes sense that the correct rendering method works better for most use-cases.

What if you try one of the GeForce virtual videocards? They feature 'full cut colorkeying' which means no partially colorkeyed pixels should appear anytime, but in turn, colorkeyed objects are 'thinner' and jaggier.

Thanks, that did the trick - results are identical to the "previous versions" screenshot in my last post.

[lowenz]

Dege, about the (possible) future of dgVoodoo2: viewtopic.php?p=662067#p662067

[Dege]

Sigh... D3D8.1 level through DX9 interfaces...

Thanks for calling my attention to that!

[lowenz]

Any chance to see in the next WIP the compatibility with the new ReShade (3.2.x) ?