Old games appear differently when played on Windows XP versus Windows 10 at identical resolutions.
Old games appear differently when played on Windows XP versus Windows 10 at identical resolutions.
I started Medal of Honor: Allied Assault on my current PC and the Windows 98 system, exploring the first level. There doesn’t seem to be any noticeable improvement in performance between the two systems, though I did observe some problems with seams on certain 3D objects on my Windows 98 machine. For comparison, my Windows 98 PC is running a Radeon 7500.
If you examine closely, comparing the bottom picture to the top one, you'll notice the bottom image appears more muted and the rifle darker. What do you mean by a practical difference? Essentially, modern Windows versions suffice for most users, but not for me. As a perfectionist, I dislike games that look altered from the developers' original design. These subtle variations are likely why some enthusiasts opt for retro gaming PCs.
If we plan to share screenshots here, it would help if people don’t tag them right away. This is the most effective method to reduce the placebo effect without doing full A/B tests.
There's the flaw that I wasn't in the exact same spot, so I can go run this again and transfer my save files over. Plus that makes it easier to run a filter in an image editor to see where the differences actually are, if any.
And I realized it now but I used different screenshot methods. On Windows 98 it let me do Print Screen, but Windows 11 had issues, but I found the game has something that takes screenshots, but in TGA format (whatever that does)
Also Windows 98 doesn't have compatibility features. I believe that was introduced in Windows XP, if not Vista.
The point I'm trying to make isn't which one looks better, it's how much of a difference there is between the two. So when I say "practical difference", I'm saying for example an RGB value of say 128, 54, 92 has no practical difference to 128,
55
, 92. Yes there is a difference, but you're not going to really notice it.
So I went back to MOHAA, made saves so that I could have a consistent position from where to take screenshots from. I also noticed that in my previous attempt, I left image sharpening on the driver settings, so I turned that off. And to make sure I'm using the same screenshot capture method, I used the in-game one.
Here's a link to the album:
https://imgur.com/a/ty0H0wl
View: https://imgur.com/a/ty0H0wl
. The album includes:
3 different scenes
Render from the RTX 4070 Ti with no compatibility mode
Render from the RTX 4070 Ti with Windows XP SP3 compatibility mode
Render from the Radeon 7500
Difference filter between the 4070 Ti without compatibility mode enabled and the 7500
Difference filter between the 4070 Ti with compatibility mode enabled and the 7500
Difference filter between the 4070 Ti with and without compatibility mode enabled
The difference filter pretty much makes it conclusive that there's very little difference between the rendering modes, especially with compatibility mode enabled and disabled. Note that the sky box in the scenes are still moving and in the third image, the NPC moves as well, so they shouldn't really be considered in the comparison.
The only reason why I didn't do a difference filter on one of the scenes is because I couldn't get a consistent image, but I wanted to include it anyway just to have another scene.
When viewing the difference filter image between the 4070 Ti with compatibility mode active and the 7500, with brightness adjusted, the variations become apparent in the same regions I observed earlier. I noticed the top left window appears less sharp compared to the previous image, and indeed both the rifle details and these differences are clearly emphasized. This suggests a distinction between older and newer video cards. Although subtle, there is a noticeable contrast, and if I could detect them without realizing they existed before your post, it indicates others can too. Compatibility mode has no impact on the visual output. This implies that GPU makers prioritize rendering new games over preserving the appearance of older titles.
What I really notice between the images is just a technical variation. I don't perceive a noticeable difference that would matter practically. Although I could grab more screen captures and conduct a blind test, most of this task was aimed at confirming there wasn't a significant change in the output as you pointed out in the forum thread.