This isn’t a tool to make good looking characters. You need the character art to start with, then this just applies a samey-aesthetic filter over the top of it which rerenders the whole scene from scratch, removing all the artistic intent behind everything visible.
Those graphical lightning tools actually helped artists create a more immersive image, with light balancing more like it was expected to instead of painted on static shadows of old.
This ai filter shit ignores light sources, purposely highlighted points of interest, and whatever else was painstakingly worked on.
And as for not gatekeeping… Just running this requires a second graphics card, at a time when people can’t afford/find one… Because of AI slop centers
This isn’t a tool to make good looking characters. You need the character art to start with
No, you literally don’t.
If character looks isn’t a concern (like random generic NPC’s in an RPG walking the street), you don’t need anything… just a blank face and let the AI fill it in with anything.
How many times do you see the SAME person in something like Cyberpunk?
This solves this, without the need for an artist to create 1000 different faces.
Do you actually believe that DLSS 5’s graphics are good looking? Lighting is completely lost as every character has three point lighting applied to them. Same with surfaces, original lighting is just gone.
Every character looks like they have a pound of makeup on their face. Details get muddled or lost entirely, replaced with what the model decides it should look like instead of the artist’s original intent.
Not to mention hallucinations and other inaccuracies. DLSS 5 uses a 2D frame for its filters. Depth and context are lost entirely as it physically cannot keep track of those things. In images with small, repeated patterns (such as a net), the model completely eats shit and doesn’t know what to do. All on top of ridiculous hardware requirements to even pull it off in the first place.
This is not the future. This is not giving people tools to make better games.
Yeah, you didn’t seem them all bitching when physics engines and game engines started doing all the stuff like lighting, shading, animation, etc.
You know, all the stuff that took an ‘artist’ and the special artist talents to create and look good and be all ‘artisty’
But now there’s similar tools for creating good looking characters they are all of a sudden up in arms?
AI won’t replace well made games or art direction…
It just stops the gatekeeping, and allows more people a chance to make a good narratives, interesting worlds, or complex game-play.
This isn’t a tool to make good looking characters. You need the character art to start with, then this just applies a samey-aesthetic filter over the top of it which rerenders the whole scene from scratch, removing all the artistic intent behind everything visible.
Those graphical lightning tools actually helped artists create a more immersive image, with light balancing more like it was expected to instead of painted on static shadows of old.
This ai filter shit ignores light sources, purposely highlighted points of interest, and whatever else was painstakingly worked on.
And as for not gatekeeping… Just running this requires a second graphics card, at a time when people can’t afford/find one… Because of AI slop centers
No, you literally don’t.
If character looks isn’t a concern (like random generic NPC’s in an RPG walking the street), you don’t need anything… just a blank face and let the AI fill it in with anything.
How many times do you see the SAME person in something like Cyberpunk?
This solves this, without the need for an artist to create 1000 different faces.
Yes, that is a thing I said.
Do you actually believe that DLSS 5’s graphics are good looking? Lighting is completely lost as every character has three point lighting applied to them. Same with surfaces, original lighting is just gone.
Every character looks like they have a pound of makeup on their face. Details get muddled or lost entirely, replaced with what the model decides it should look like instead of the artist’s original intent.
Not to mention hallucinations and other inaccuracies. DLSS 5 uses a 2D frame for its filters. Depth and context are lost entirely as it physically cannot keep track of those things. In images with small, repeated patterns (such as a net), the model completely eats shit and doesn’t know what to do. All on top of ridiculous hardware requirements to even pull it off in the first place.
This is not the future. This is not giving people tools to make better games.
yeah well, all you naysayers said the same when DLSS 1.0 came out…
Whatabout