Paint Tiles

Ok, thanks! I think I am going to try using the Apollo palette to start, slightly larger than DB32, Using my drawing tablet.
(46 vs. 32 colors.) but it's well put together too, and has a good range of hues.

I have debated maybe experimenting with a more 'painterly' approach to palette though for something like this. 
Going down to a really small set of colors which correspond to particular paint types. but allowing mixing between them. (Which I guess technically gives a much wider set of colors even if the range is somewhat constrained?)

As practically it would take some sort of custom paint mixing, and then trying to visually match to the palette colors to actually match a hexcode based palette with physical paints.

If I wasn't trying to be as colorful. I have worked with the zorn palette before. But it likely would have a pretty different effect.

If I wanted a stronger blue I could do maybe the Reily Palette.

(or maybe go for a palette with no black that uses blue-red mixes for near blacks?) 

Although, for effective usage as tiles I might end up just having to effectively build my own broader palette out of a small number of source  colors anyways. (to get the actual digital form?)

And the question of how many in-between mixes I use is significant.

In my digital experimentation, one thing I have noticed makes things pop more is to use two similar or adjacent colors. putting one below the other and leaving some textural holes in the top one to imply some degree of shading. not sure how well that would work with 32 colors. But to an extent batching colors like that might allow me to reduce the total number of tiles and get a bit more depth to an extent. it does basically reduce each color into 3 lightness levels though. 
It does noticeably have some cases where a color fulfills a role across multiple ramps to be more efficient. 
It only really makes since for larger block colors, any line work is going to want most of them.  Or at least all the very light and dark colors.

Dawnbringer 32 is pretty brown and yellow heavy. most colors have 4-7 slots. (A few if counted sparingly are only like 3.)
and like 9-10 browns. depending on how you count. (you could throw some on a seperate yellow ramp though.)


As if I actually painted them out on canvas(Or more likely tiny wood panels. As they tend to be thinner? Probably would go for something like 12x12 inches or maybe 6x6 inches.(for the actual tiles but maybe do like 4 per panel? which might go as small as 3x3?))  doing 4 per panel might make digitization and being cohesive easier, But would make them less useful as actual physical tools?

I guess I could get one of those cheap packs of 3 in or 1 inch canvases. But I suspect it would be hard to do much other than a solid block or really thin line work that small. 

Not sure if I will or not. (Might have a cool element where I could assemble them somehow to make bigger works temporarily.) 

Only issue is that I wouldn't really be able to apply the layers unless I baked them into a tile? 

Unless maybe there is a medium/ground which is both durable and allows some degree of transparency? Would need a mounting setup probably then? 

Like some forms of paper are transparent? I could also  cut out a form? Maybe using something like cardboard or similar. If I wanted to apply a curved line as a separate piece? 

But how to effectively mount it without damaging the underlying tile and being removable would be important?

I guess maybe I could make the paint extra thick, and throw some sort of protective thing on it and stack vertically? 

So rather than hanging you place it on a small table and stack the other pieces? (Would probably guarantee some damage though?)

Maybe I could use foam on the back of the pieces to give them more thickness and potentially reduce the damage they might do?

Oh, maybe if I wanted to get fancy I could use magnets? 

Lot's of possibilities, want to start with something simpler though. 

I am currently working with a digital painting program for convenience, but maybe I could pull out my actual oil paints I have? so long as I could figure out how to effectively digitize them without having the filesize explode.

I am sort of curious how  you decided which tile size to go with.

Another idea I have considered is applying the same basic pattern at multiple resolutions, to allow fitting characters composed of multiple squares into a larger whole.
(perhaps 200x200 vs. 100x100.) Although, from some of my experiments, some of the simpler larger blocks still look pretty good at 50x50. A lot of the patterns stop feeling painterly there unless the fill an entire  square. (Even like a right triangle with short sides at 50 sort of is a bit too small for a lot of them to work.)

Part of that is most of the painterly brushes I found only really have texture if at least 20-25 pixels in size, which makes sense.

So practically, using a larger one as a base and just going bigger makes sense 
But at 1080p 100 pixel tiles are already around 1/10th the screen.
Although, if they have 4k? then a similar ratio could be retained. with 200x200 tiles.
1/7-8 is probably the lowest I would want for gameplay purposes assuming no 'tall walls' or platformer like levels.

1080/8 gives about a maximum of 135 pixels.
I guess I could try looking at 67 or 68 pixels as a base? Which would let me double to about that size? Maybe the extra 18 pixels would be enough to make patterns feel nice.
(Or could make it 64 if I wanted to be a nice multiple of  common pixel art sizes.) But 14 extra pixels might not be enough? (I guess I might be able to fit 3 thicker textured lines. so presumably less textured hatching patterns would still work. If each is a about 20, Maybe I could fit 5 thinner lines with a good looking spacing and variation? 

I sort of realized why you probably but 20 pixels of bleed. as that let's you draw lines along tile edges which display over the edge of both tiles if they are spaced 100 appart, so have that much overlap. 

I know I kind of wrote a lot, sorry if it's too much. just excited.

that makes sense.
This is such a cool idea.
If you can remap colors, then to an extent the programming logic, and art time can be spent making a larger variety of textures, rather than say repainting the same shape in a bunch of different tones.

Part of me wonders how feasible it would be to use  something like this in a procedurally generated level. Or if it would basically mandate hand-done ones? 
Regardless the tile placement really would need to work in at least 3-5 layers. Unless you want it to explode outwards in terms of  predifined files. As most spaces are going to have at least 2 blocks of colors. and then possibly a line and a pattern on top. (Or multiple.)
And if you want things to feel connected then you need to have that encoded somewhere. Almost seems like it's full potential might require something like waveform collapse

Thanks for reply! In this case, I guess the solution should be to separate the tools and resources. Users can either purchase your resources to import or create their own resources.

I am using this tileset in a commercial game and plan to ship a level editor, but allowing for full customization of the level background would be redistribution as you've defined. Having the user import a resource wouldn't quite work either, as the game relies on a modified version of the tileset.

Any ideas on how I can approach this? Worst case scenario I can just remove background customization from the editor, but I would definitely prefer to allow it somehow. I would be willing to explore licensing options, e.g. giving you a cut of sales or paying up front.