Dana Callista Lexa 🏳️‍⚧️

This was also my knee-jerk reaction, but somebody framed it to me more that pixel art is actually really expensive compared to using 3D models. Which is true -- look back 20 years to the Gameboy Advance, and that platform established pre-rendered sprites as something that was quick, cheap, and easy. Having a pixel artist paint a thousand frames of animation costs a lot more than just keyframing a single skeleton and then running an export script.

Like, look at Superstar Saga against Mario vs. Donkey Kong. MvDK has 2-3 times as many frames of animation as Superstar Saga, and it uses pre-rendered graphics for its sprites.

Iizuka even touches on this a little bit, noting how easy it is to attach cosmetics and change out skins entirely for a 3D model. If you wanted to give Mario there a new hat, you'd have to modify ~120 individual frames of animation, one by one. Whereas with a 3D model, you just remove the old vertexes and attach the new ones.

This is less "pixel art as a medium will be dead in ten years" and more "2D animation as a whole is less cost-effective than 3D animation." It's the same problem as Disney ditching ink and paint for CGI (it's just framed with pixel art, given Sonic Mania is still a fresh memory).

All of this is to say: it sucks, because 2D animation looks wonderful, pixel art looks wonderful, and these people need to stop thinking so much with their bank accounts. At the same time, though, as someone who is a Sonic fan and looks at basically every single game under a microscope, I'm not alone in saying the last 3-4 Sonic releases have felt very low budget. Even (and maybe especially) Sonic Frontiers. And Takashi Iizuka, despite being the head of Sonic Team these days, has spelled it out that he's still beholden to Sega's corporate whims and isn't actually allowed to make the games he wants the team to make.

Ergo, I can imagine the dude is looking to save a few bucks wherever he can, because he has to work with what he's given.

I'm speaking here as:

• Someone who has professionally done pixel art for games
• Someone who has professionally done rigging and 3d animation for games

"How expensive" an art style is is heavily, heavily depending on the production in question and its needs. Not just "are there alternate cosmetics" but stuff like:

• Do you need alternate angles, and how many?
• What kind of frame count are you going for for standard animations?
• How much bespoke art is in the game vs how much standardized art?

Generally speaking, 3d is very good at minor variations on a theme. Retextures, small outfit additions. 3d is very bad at bespoke additions. Pixel art, meanwhile, is expensive for very large, complex animated pieces and variations on large-frame-count anims but VERY cheap for bespoke additions.

Things you can do very cheaply in 3d art:

• Add props to existing characters
• Retexture existing characters
• Create new model variants for existing characters - as long as the body type is similar or identical
• Vary animations procedurally
• Have physics interactions
• Have many camera angles

Things you can do very cheaply in 2d art:

• Create wholly new assets/characters/animations extremely quickly
• Create sequences that would be complex and difficult to animate in 3d, like complex prop interactions, character-linked effects, etc
• Create games with a minimum of staff (the cost bar for pixel art is lower than 3d art at the low-budget end, due to 3d art's large wind-up cost - making and animating a single model requires at minimum 3 different disciplines. That sunk cost becomes lesser on larger productions that need huge numbers of assets per character)
• Create animations without worrying whether body mechanics are an issue or dealing with technical know-how for issues like gimbal lock

They're different mediums and they have different requirements and strengths. There is no hard rule about which one is going to be pricier, it's 100% needs-based.

I don't want to shoot the dog, but I do want to shoot the idea of having a dog in the game to pet.

I know, I know, it's pretty mean of me to pull out a gun and unload a full clip into a completely adorable and wonderfully intentioned little trend even if it IS a little bit annoying, and it's certainly kind of an overreaction, but while I've got this gun out and you're listening would you mind if I made a particular hyperspecific point? I take issue with the idea of putting a dog in the game to pet because I think that's missing the point: if you have a dog in a game, you should be able to pet it.

Do you understand the distinction? The only reason for asking "Can You Pet The Dog?" is because the dog, when present, doesn't give the player any way to interact other than what the player normally has, usually gun. Gun is not the ideal way to interact with dog, and developers tend to forget pet but remember gun because petting a dog doesn't do anything. For player's that's exactly the whole point, that there is no point, and that's kind of mindblowing to me. Like, I spend all day every day trying to make a big digital stage and trick people into believing it's real, and they laugh at your antialiasing and say you need to upgrade your version of the Unreal Engine, and then they'll go on to ask why they can't pet a dog. Isn't that funny? They think they see all the fakeness and seams but at the end of the day the human brain is not immune to seeing a picture of a dog and going " doggie :)"

My best friend and streaming partner Fern loves Bloodborne and Souls games more than anything, and one of her favorite things about these games is the ability to wave and gesture at people. They put emotes in those games because of multiplayer, but you can do them whenever you want. Wave to the Silver Knight shooting arrows at the pillar right in front of you. Wave to the boss through the fog door. Wave to the fire keeper—wait she waved back!

You can't "pet the dog" in Bloodborne (dogs in these games are made of knives, unsafe to pet) but you can "pet the dog" (the doll giggles when you act goofy, like a teenager trying to get a girl to notice her). This fake, inanimate world responds back to the player's attempt to communicate with it, even when there is absolutely no mechanical reason to do, an act of communication between player and world that is delighting in and of itself.

This is the spirit, rather than letter, of petting the dog. It's not about a literal dog that you literally pet. You can find a way to put a dog in the game and you can figure out a way to pet it, but it's not very impressive. What I would like to see is identifying what in the game the player wishes they could interact with, simply for its own sake, just because they believe in the game, believe in it more than you who made it does, and want the game to speak back in a way that validates that belief. That is a much harder dog to pet. But isn't the whole point a lot of work for no real point?

Speaking of, while I've got this gun, I feel the exact same way about fishing minigames. A calming and idle pastime in the middle of a very different kind of game is a great idea, but I think you should pick one that suits what your game is about. Yes, Cruelty Squad does have fishing in it, but the real fishing in Cruelty Squad is the stock market. Think about that while I lower my gun and give you a chance to flee.

there's so many fucking things that piss me off, but one of the things that pisses me off the most is basically every tech-focused indie dev who sees a shiny new toy (raytracing, nanite in ue5, DOTS/ECS in unity) and then apparently immediately suffer a blow to the head and completely forget about how people have made games for fucking decades

like for some reason a lot of unity devs are currently CONVINCED you just can't make games with the aforementioned tools. every time somebody asks how to do something, down to how to implement systems that were used minecraft, they do that thing like they're a stereotypical car mechanic and go "welllll it looks like you're going to have to implement a robust entity component system based architecture, gonna be pretty hard to learn" as if minecraft hasn't been around for nearly 12 goddamn years at this point

and like somebody else asked how to check if a player was in shadow in a game, something you can do in a particularly naive implementation just by casting rays from the player to each light, y'know? but somebody was like "okay so you're gonna want to use command buffers to blit a shadow map to a custom render texture and sample from that to see what the light level is there" like what the fuck? like this isn't a new tech thing but it's one of those dramatically overengineered solutions that drives me nuts

and there's a much of people who are like "unity is GARBAGE if they don't implement nanite. I NEED nanite to make my game" as if every other engine out there has been using virtualized geometry for a billion years and it's not just something that became generally available in ue5 for a bit over a fucking year while they've got a bunch of WIP screenshots and they're making a fuckin' 3d puzzle platformer

I understand the allure of shiny new tech, but people need to fucking chill and think about how we actually have been making videogames for fucking ages

a really big part of the problem on a fundamental level is that gamedev tooling is exceptionally poorly supported and exceptionally secretive.

This industry has had literally decades to develop a robust set of gamedev tools but because we are

• completely insistent on reinventing the wheel absolutely every time
• impossibly secretive over the most basic features
• completely unwilling to fund, use, or support open source tools
  we end up with situations where - I've been working in games for literally 12 years and nearly every game I've worked on has reimplmented some fundamental thing like dialogue systems from scratch.

Can you imagine how film would have ended up if everyone had been like no I can't tell you how to move the camera nicely it's a SECRET

But that's basically where we are, in this embarrassing situation where gamedev just hates itself so thoroughly as an industry that we have no institutional foundation and even the 'easy solutions' we point new devs at still require them to build them from more-or-less scratch

the vast amounts of serotonin that can be gained from making my funny gal move around on screen

(gamedev listening in the background; That's Just It - 2 Mello )