Graphic Imagery

This week of development was spent split between attempting to instill greater ramifications of purpose into my virtual world and mere cosmetics.

Starting with the cosmetics, I must be serious about making my game, because I spent more than the price of a AAA game on those cosmetics.  I bought two tilemap packs from Oryx Design Lab: the Ultimate Roguelike Tileset and the 16-Bit-Fantasy Tileset.  Of course, such a price is a heckova lot cheaper than hiring an artist to make your own.

I had hoped that the Ultimate Roguelike Tileset would have worked, because I wanted a tileset with a western-theme and a black background to serve as a slate for the imagination.
The sample image provided for the Ultimate Roguelike Tileset certainly looks nice, but you might notice something a tad lacking in uniformity when you compare the spacing of the dragon from the wall on the far right and the spiders to the left (but relative to the left wall or the bridge to the south).  Well, it is just a mockup, after all, but I do have to wonder at the viability of the set for procedural generation purposes when even the mockup cannot seem to keep its tiles spaced evenly.
These purely one-color tiles have a significant advantage when loaded into something like Unity because I can, on the fly via code or editor properties, re-color or re-scale the same sprite and make it represent different, distinct things.  For example, I took the same tree sprite and had it represent a light forest when thin and light green, and a heavy forest tile when fat and darker green.  With such rudimentary graphics, the potential for polymorphic transmutations and easy modifications on behalf of the developer is quite high!

Unfortunately, I encountered a few problems with it that I decided made it a wasted purchase on my part:
  • My game design currently calls for large wilderness areas, including mountains.  I can't use the v1.0 for this very well because, if a player approaches a mountain from the left, right, or back, then there's not really identifiable mountain tiles.  v2.0 has more parts and probably could provide a better experience, but unfortunately I have no idea how the tiles fit together. 
    • Mailing the developer, I'm not entirely sure he does either.  He suggested I load up an included Tiled file example.  There is none.  I guess he's so busy he forgot which of his tilesets had that and which didn't.  So I imported the tiles into Tiled and experimented with seeing if I could figure out how the v2.0 mountain tiles fit together.  I still have not figured it out.  When your tileset is made up of rudimentary lines, documentation becomes more important.
  • I wanted an old school PC look, but perhaps this is a little too oldschool.  It is downright Amstrad in places.
  • I thought I could get Unity to handle the spacing of tiles that are not of equal height and width (after all, Caves of Qud does it just fine), but ultimately it proved a little beyond my technical skills.  I decided it was more important that Unity editor game space corresponded to the coordinates of the tilemaps it was displaying.  The only way to make the tiles match with a 16x24 tile is to scale the X scale of the sprite by 1.5, but this naturally made the tiles look a little uglier.
Now, we do know for certain this tileset can work.  Awhile back they ran a competition, The Trials Of Oryx, where they provided the tileset along with a cash reward for best game for the competition.  It was a clever move, the result was indeed many examples of quite-functional games that use the tileset.  To see a bit of the kind of trouble I was talking about, a good example of such a game would be the 5th runner up, Tull, playable right from the web.  
Cropped Tull screenshot.  The player character near the bottom is distinct enough, but the magic shop merchant's coloration fairly blends in with his tile, and you can see bits of it piercing his character model.  It is not entirely the tileset's fault it was employed this way, but it does show a way in which it cannot be employed without getting a little too old school.
Tull uses the v1.0 of the Ultimate Roguelike Tileset to fairly good effect to make the kind of game I want to make.  The rudimentary nature of the graphics does not "pop" nearly as well as the mockup.  In places, the characters bleed against their flooring.  This is a problem with this extremely old school style.  If you look at the Ultima series, the first five games, you will notice Garriot and company chose to have characters completely black out the tile behind them, even after they were using platforms where this was no longer a technical requirement.

Overall, I can't really blame the Ultimate Roguelike Tileset.  Honestly, there's enough great parts in here that you can fudge it even if it doesn't give you exactly what you want right out of the box.  To an extent, I have to blame my lack of perseverance or creativity for being unwilling to do so.  But I prefer to think I just wanted some graphics that were a bit more distinct when layered and of equal proportions on the x and y axis.

Now, there was the Dawnlike Tileset but, for all of its free and extremely diverse loveliness, 16x16 size tiles were just slightly too fuzzy and indistinct for my liking.  Thus, a great selling point for me to give Oryx a second chance was that the 16-Bit Fantasy Tileset had fully 24x24 size tiles, 150% the fidelity.
Now, I know the 16-Bit Fantasy Tileset will build correctly because he built these mockups in Tiled, and the graphics are detailed enough that I can pretty much implicitly figure out where each tile goes right from the get go.

I can highly recommend this tileset, as it is well-designed and fairly easy to implement.  Unlike the current version of the Ultimate Roguelike Tileset, it includes the original photoshop files as well as a Tiled scene that makes it easier to understand how everything fits together.  (Although I will say that the Dawnlike set is a bit more explicit in its instructions, providing a nice diagram on the top of each page to help figure it out, albeit a bit like decoding hieroglyphics.)

All this cosmetic distraction was basically my psyche finding escape from the far more difficult task I've been trying to confront.  The goal of my procedural generated, turn-based roleplaying game is to have a persistent state tile-based world where everything has a purpose; nothing is generated without a higher significance attached.  It is a tall order, but also a mandatory feature, and I am currently directly confronting how to go about doing this.

I had hoped to get that figured out during the four days of this bizarro weekend, but it seems I have run out of time with very little progress made along these lines.  In regards to solving that problem, I seem to have landed back in old trains of thought.  It really is something I feel that the game developers of the world ought to have made but now, but I suppose it's up to me.

I will share a strong point that I find has helped a lot with making progress: it's not about finding the perfect or right answer, it's about finding an answer that is more than adequate for the task.  This sets the bar in a far more reasonable place for tackling what seems impossible, as many solutions I have stumbled upon are surely more than adequate, even if they are not without flaw.


Popular Posts