Hi, newbie here. I just look at this kind of images and I would like to have a game with this kind of feeling. Except for the pixel-art 2D part, which i understand, how could i get lights like that?
This could be 100% in 2D or it might be an ortho 3D view.
For example, note that there are many tricks going on above.
The most obvious one is the illusion of three point lights (the sconces). If there were three lights then those shadows would be staggered in shades of gray, not sharp and black like that.
There’s also some fakery on the tops of the pillars: note how the smaller block atop each pillar, while it casts a shadow behind it, does not have the shadow of the pillar itself drawn. This might just be because the real light is above the height of the pillar, if it is 3D.
In fact, looking at some of the finer details, I think a lot of it is either painted into the geometry, or else the entire thing is just a photoshop, perhaps a marketing image made from a realtime game image that didn’t have any lighting at all, or perhaps far less interesting lighting.
Remember: it’s all tricks you kinda have to noodle with yourself.
EDIT: don’t get me wrong, I think it looks amazing, and at speed it would be glorious. Just know that almost everything in gamedev is some form of Potemkin Village.… that’s just the reality of realtime software rendering on consumer-grade hardware.
As stated by the author on the pinterest page, this is using 3d lights that dont effect the normals[1]. As far as I can tell, the only different thing here is only 1 light creates the object’s shadow. This pillar’s shadow bending from objects are easily obtained with 3D objects + 3d lights
You could actually get a similar look with 2d lights + 2d sprites too[2], but it is way too hard for you to even worry about. In short, you’ll only cast 2d shadows to the ground layer, use the resulting “Light Texture” to add darkness effect based on a secondary texture(kind of a height texture) of the sprite. For shaped shadows like the one on the police officer, you’ll use the negative lighting technique[3] based around the strongest or closest 2d light source.(this image doesn’t have negative lighting)
[1] it could also be that the 3d lights effect the normals of 3d objects, but the characters are 2D(using a plane–which wont really have normal information)
[2] you can’t feasibly get fine details such as the shadows to stop based on the caster’s height vs light’s height like this:
[3] you can see the negative lighting technique in the happy harvest demo