In other news, I began writing a new huge blog piece about the Holy Trinity of Game Design (aka how to be sure of your game design).
It will explain that there is indeed a holy grail of sorts for making a perfect game design, regardless of the project and its intentions. Will post it both here and on gamasutra once its done, but it may take a few weeks before its ready.
Either way, it’ll be big.
Nukes. Bombs. Missiles.
Terrifying things in real life, but awesome in a virtual reality.
I’m am noticing that even with a mass production framework in place, making pixel art according to a strict scale and realism standard is still quite a slow process. I remember back when I made mods, making graphics for them was a matter of minutes and I could generate a lot of them in a single. Then again, they only had to adhere to a limited color palette and there were plenty of templates to give a clear idea of how to start drawing the sprites. Also, since the mods as a project weren’t all that serious, I could get away with cutting corners with measuring the scale or caring at all for realism/consistency/shading.
Thats not the case with the PSRC project sprites. Each object is like a small puzzle that needs care for it to not be useless once it done. Theres no point in drawing a sprite for a library if its scale is measured wrong or the shading is poor. You simply wouldn’t be able to use it along side all the other sprites, since it would either ruin immersion, feel wierd/out of place or have a critical functional flaw that prevents it from being usable with other sprites in the library.
I want to make a complete tutorial on how to make sprites sometime, but the general steps to create a PSRC sprite are these:
1. Find a reference
This must be either a photo of the object, or atleast a realistic drawing of one. Bonus points if it also has its measurements on the image to give an idea of how big its really supposed to be. Also the best case if its pictured directly sideways and does not have a black background behind it. Neutral lighting to make sense of the details is nice too.
I usually end up spending plenty of time on the internet or anywhere else I can find images to hoard a large collection of references before actually drawing anything.
2. Measure and figure out the size of the object
This is tricky as drawing with pixels are not as freeform as drawing with a pencil on paper is. You have a limited amount of pixels that you have at your disposal and each of them has to count. They also need to be all placed manually by hand. I don’t do resizes of reference images, since it looks shitty and is much more difficult to work with.
While I have the benchmark of 40px = 1 meter, theres a small caveat there. The average human is only 60 pixels tall, which is 1,5 meters. This means any objects that I draw should also be 20-25% smaller than what the reference might say, otherwise you’d have objects too big for the human to interact with and make an adult look like a small child. I’m happy I noticed this before releasing the first wave of packs.
During measuring, I’m also looking at what detail to keep and what to ignore. Pixel art, given what it is, its often impossible to include every tiny detail on an object, so I need to use some artistic license to make sure the object looks right, but not be too crowded/messy with too much detail attempted to be crammed in a tiny space.
I also have to verify that the object is actually realistic/believable, as if its not, it’d be like having a silly cartoon character in the middle of a realistic environment. Out of place and impractical.
3. Blobbing it out
As seen in the top picture of the post, thats what I mean by “blobbing”. Using only 3 shades of grey, I create the correct shape of the object and its major details.
Ideally, this is the last moment I want to make sure that the size of the object is correct, because doing it in the shading phase is not as easy without having to reshade all of it.
Since all I’m using here are three colors, it may feel difficult to see what part is where and how a certain detail is supposed to look like in the finished version, so I need to keep a mental bookmark of where each detail is and use whatever means I can to not get confused here. Using more colors and drawing with outlines are some of the tricks I use to figure this out.
To make sure the size is correct, I can copy paste a human character template and hover it as a selection near the object to make sure the size checks out.
Hopefully by now I’ve made sure the scale is right, the proportions are correct, the details are in and the margin of human scale error is taken into account, I can then finally begin shading.
I need to decide what the materials are that the object is made from, which will mean a big difference in how it will be shaded, since shading is basically how the object reflects light off it. Is it shiny, dull, reflective or transparent? Those are the questions I need to answer. Also if it has a texture or not. For metal, shading is enough, but for wood, I might want to apply a texture to make it actually look like wood and not plastic.
I sometimes consult a video or a guide image to refresh my knowledge on the shading stuff, as there is a lot to remember if art isn’t the only thing you’re doing.
I’m using a burn brush with very specific settings to lighten/darken the pixels to apply the shading to an object, where I make sure that the shading stays within a certain range of values, giving the sprites a crisp look, while avoiding to use too smooth gradients to blur the image into mush.
This part is fairly easy and freeform. When the greyscale image is done, I can use it freely to apply any color or any pattern very quickly and easily. All I need to do is to make a new layer set to Hard Light mode and then just paint the single color over whatever parts I wish to be that color. I don’t need to concern myself with darkening or brightening the color, since the shading of the greyscale image does that automatically for me. This is why I mention in the promotional images that the sprites I do are easily customizable and flexible.
6. Clean up
To finish up a sprite, I need to make sure there are no rogue pixels left from the background I was drawing on, or otherwise you’d have a random pixel sticking out of a sprite like a sore thumb, effectively ruining it.
After that, I crop the image and save a greyscale version and a default colored version of the sprite, which are then finally ready for consumption.
All of this is hard work to do and takes a long time, but it only needs to be done once for each time. Unlike many other pixel are resources out there, these work kind of like realistic 3d models, except in 2d form, that you can repurpose in any project and have it work nicely with any and all other sprites that are part of the library, that will eventually cover anything from apples to fridges to nukes and beyond.
BTW: I just realized I had titled all my march posts during april. Whoops. Guess april doesn’t exist on my calendar. Skipping straight to may lol.