Am i the only one who hates the use of dithering in modern games?

interestingly yes.

the PS2 was so good at layering transparencies on top of transparencies that porting games to other systems could be problematic.
the Xbox version of MGS2 runs worse than the PS2 version, despite the Xbox's GPU being more than 3 times as powerful in raw compute numbers... and the reason were transparency effects like rain and layered textures.
japanese devs struggled with the PS360 gen for similar reasons. they were used to just layering transparent layers on top of eachother, which worked on PS2, but completely destroyed perfomance on PS3 and 360.

the way many engines do transparent effects now btw is to use dithering and then smear the image to make it look transparent... yes, just like the fucking MegaDrive did 🤣 but now they do it through TAA instead of the CRT blur.
that's why many modern unreal engine games look broken if you turn off TAA, and why they usually don't let you.

Nintendo usually doesn't use TAA, and so they can't just smear the image to hide when they do perfomance saving measures like dithering. but in return you get a cleaner image that isn't butchered by TAA blur, TAA artifacts, and TAA motion trails

PS2 was a beast!
 
Another one of these artifacts that were introduced when we went to PBR?

For a lot of modern games they expect you to use TAA, AI upscaling, or whatever, to smear away the artefacts.

Plus resolutions above our loved 1080p
 
Last edited:
I don't mind it at all.
The higher the resolution, the finer the dither.
It's only when done on very low resolutions, its a bit much.
I found it very annoying in Neogeo games. Sega genesis did it to, but it had some charm.

Extreme visual quality
optimized for good performance
High resolution
Low cost hardware requirements

You can pick 3
 
Last edited:
I understand why it exists, but I find it extremely noticeable outside of transparency got full objects.

Capcom love using it for shadows and it looks bad. I know shadows aren't cheap to render but I'd prefer the choice for an alternative because there's a line and that crosses it.
 
Last edited:
interestingly yes.

the PS2 was so good at layering transparencies on top of transparencies that porting games to other systems could be problematic.
the Xbox version of MGS2 runs worse than the PS2 version, despite the Xbox's GPU being more than 3 times as powerful in raw compute numbers... and the reason were transparency effects like rain and layered textures.
japanese devs struggled with the PS360 gen for similar reasons. they were used to just layering transparent layers on top of eachother, which worked on PS2, but completely destroyed perfomance on PS3 and 360.

the way many engines do transparent effects now btw is to use dithering and then smear the image to make it look transparent... yes, just like the fucking MegaDrive did 🤣 but now they do it through TAA instead of the CRT blur.
that's why many modern unreal engine games look broken if you turn off TAA, and why they usually don't let you.

Nintendo usually doesn't use TAA, and so they can't just smear the image to hide when they do perfomance saving measures like dithering. but in return you get a cleaner image that isn't butchered by TAA blur, TAA artifacts, and TAA motion trails
You are over simplifying there when comparing generational leading fragment capabilities of the PS2 with the PS360 gen hardware.

What changed is that the tightly controlled camera design of the halfway house isometric-2D/3D PS2 gen moved to a more generalized Mario Sunshine full 3D camera setup across the board meaning that many of the deterministic and redundant workloads in games had become non-deterministic and needed recalculated every frame and despite the far more powerful hardware, the step up to 720p and generalized alpha workload could exhaust fragment shading without a different algorithmic approach or sticking with a isometric 2D/3D halfway house -like many Japanese devs opted for.

Despite Mario 64 felling completely 3D, Mario Sunshine was chosen by me because it represents a big step up in its camera freedom with lots of water fx relevant to the topic and even uses stencil hatching for partial transparency in many areas too mostly for artistic control of the final scene composition.
 
Remember the Genesis? It used dithering to blend it to fake colors and transparencies on old TVs using composite.

Remember the Saturn? And how it also used dithering because it was hard to program real transparencies because of a hardware limitation?

It's good that modern systems have no such limitations eh?

So why this?

5X0pfr5EG5vM4jz0.jpg
OahGBruTIUxyZGJv.jpg


A lot of Nintendo games seem to use this effect lately but other devs use it to dither things that get in the way of the camera, mostly foliage, for example The Witcher 3:

fz1LuwBeqkn9hdD8.jpg


I remember playing a bunch of newer games where this happens, especially in stealth sections where you hide behind stuff and that stuff gets dithered.... But i don't remember the titles at the moment.

My point is.... Why even use this archaic technique on modern games? We used to get real transparencies for this effect so what happened?
DLSS or upscaler can cause this if I'm not wrong. Maybe also low buffer artifacts. The price of 60 fps I guess.
 
Last edited:
it only has to draw and calculate HALF the transparent pixels. transparency has to be calculated for each pixel.
It doesn't even need to do transparency at all. It simply doesn't render the object on the alternate pixels.
 
Last edited:
PS2 was a bandwidth beast. Everything that came after is just an underwhelming compromise. If we had kept the upscaling in amount of RAM and bandwidth the PS5Pro would need its own nuclear reactor to function. So thank you, Greenpeace.
 
Top Bottom