Comment by gfxgirl | Hacker News Reader

Comment by gfxgirl | Hacker News Reader

Parent Full thread

It's not that hard ... Or rather it's almost all solved problems. The proof is the number of 1-4 person indie teams making games that look like AAA games.

Two examples.

https://store.steampowered.com/app/417290/Ghost_of_a_Tale/

Made almost entirely by one person

https://store.steampowered.com/app/242760/The_Forest/

Small team

There are tons of others. Sure, someone had to solve those issues but they mostly packaged the solutions into Unreal, Unity, etc.

It's like concentrating on making cameras instead of making movies. Sure there are new techniques that might give your movie some special edge (like when they invented bullet time for The Matrix) but a good story (or a well designed game) doesn't really need the technical tricks. In fact some of the most famous games are designed around the limits of the tech (either the limits of the hardware or the limits of their particular engine).

Note I've written several AAA game engines and have friends that work on AAA engines for others. They are all admitting / lamenting now that engines are basically a commodity and their skills are not needed. They're turning into scripters for the artist and designers and management is pushing to just license the engines.

Teams of 1-4 people making indie games are either:

1) Using an existing product like Unity or Unreal Engine or GameMaker. 2) Using their own engine but with very basic graphics.

"Ghost of a Tale" uses Unity. "The Forest" uses Unity.

So yeah, it's not "that hard" if you use a product which has already had millions of developer hours put into it like Unity or UnrealEngine.

Take away any prepackaged game engine and see if either of those two developer teams would have been able to make those games. :)

> Take away any prepackaged game engine and see if either of those two developer teams would have been able to make those games. :)

"When you're an adult you won't always have a calculator with you !"

Why would you take away those engines? What does it matter that the games used Unity (or any other pre-made engine)? The point of a game is to make a fun experience not to reimplement every aspect of interactive software.

Do you make the same comment about someone using standard libraries or frameworks? How productive would you be without system libraries or even an OS?

BTW, I have ZERO problem with game teams using engines. If I were to start making a game now as an indie developer I'd almost certainly use Unity too.

I was more making the point that starting by building an engine is a huge commitment and you need the skills and times (and the reason) to do that.

Your original reply to the GGP comment was:

> So yeah, it's not "that hard" if you use a product which has already had millions of developer hours put into it like Unity or UnrealEngine.

What point are you making? The GGP correctly claimed it's "not hard" for small teams to make games. The games they referenced were made by small teams. The fact those games used an off the shelf game engine is immaterial. There's no "point" to make about it.

Your "point" comes off as a qualitative judgement about those developers using an off the shelf engine. If you've got no problem with someone using Unity what point are you really making?

The article is about how hard it is to productize game graphics and the GGGP said its not that hard because there are a ton of small team games with great graphics. But those teams aren't productizing graphics, they're using an engine that had millions spent on it and a huge team which did it for them. I believe that was the point being made: that its a false comparison.

Using someone else's hard work doesn't mean that its easy, just that you don't have to do that work yourself because others have done it for you.

> Your "point" comes off as a qualitative judgement about those developers using an off the shelf engine.

I didn't read it like that at all. In most cases, for most teams (certainly small ones), using an off the shelf engine is absolutely the right call, exactly because it means you don't have to solve the hard graphics problems yourself. But using an off the shelf engine does not mean that those hard problems don't exist or aren't hard, just that you can outsource them to the engine vendor.

If you make a game with Unity, you have productized the graphics. It does not matter at all that Unity has millions of dollars or thousands of person-hours built into it. The game developer paid the asking price. Everyone was compensated. The game developer didn't somehow unfairly "take" the hard work of the Unity devs.

From the point of view of a game developer, having good graphics is "easy" because Unity and Unreal exist. The fact Unreal and Unity exist enables thousands of developers to make games that couldn't otherwise. Making a "point" that a small game developer didn't write their entire stack is just shitty gatekeeping.

> From the point of view of a game developer, having good graphics is "easy" because Unity and Unreal exist.

But this is not what the article, and this comment section, is about.

You are aware that almost all of the AAA games released in the past 2 decades have used an engine of some sort?

Unity is popular with the indies, but Unreal and Frostbite have powered games with collective sales of several hundreds of billions of dollars.

Not sure if you realize it but you’re just making the same point that OP was already making.

Are there scenarios where using an engine would bbe a problem? I was under the impression most engines out there will do whatever you want to do. So I'm curious if people are still running into roadblocks because of an engine?

Depends on the idea. I started with mods and early game makers in the 1990s then prototyping in C++ with basic libraries like SDL. To make Pong it didn't matter much. To make an epic 3D FPS the engine was a do or die decision.

It was do or die, but you could pick an engine right?

There are engines that will allow you to do a 3d FPS and make it look very good without ever having to worry about any low level barycentric coordinate issues.

Isn't that a good thing?

It is good to have options. It's just that using an off-the-shelf engine means fitting your idea into that square hole. Or else you'll spend time fighting it. And for games imagined with a distinct visual style that may be a lot more or less work depending on which one you choose.

I would add that using an existing engine rendering pipeline with shaders is way easier that rolling your own. Making a small software renderer from scratch is a very good way to get a good grasp on how game engines work internally.

You are right that "off-the-shelf" engines have tooling and features rich enough that even a single person can make a pretty and fun game.

But I disagree with the higher level idea - that this means that things become simple. It's the other way around - they became so complex that a few person studio cannot write a new, general engine (see how many people programmed Doom or Quake which were state of the art back then!). At Unity there are a few thousands of just programmers working on the engine (yes!). This is not FAANG size, but I don't know if there is a team at FAANG that has a few thousands of engineers working on a single product. You need a team of this size to make a multi-platform, universal engine that is still not good enough to make AAA or open-world games.

Also my personal experience at my last games job (I left games) was the complete opposite of what you suggest as "engines being commodities" (at least this is at odds with you writing that they are working on AAA engines- sounds like they might be "using" engines and not writing in-house ones?). At Sony Santa Monica we had a team of ~30 programmers, out of which ~8 graphics people (just for a single platform and a single game) and I was constantly frustrated with how impossible it was just to catch up on all necessary state of the art techniques with a team of that size.

There are people much smarter than me who spend for example a half of a year on something as obscure as "multiscattering rough diffuse BRDF" - and a larger engine has hundreds of "features" like this.

Things become simple by abstracting out and hiding the complexity.

My teenager doesn't need to know about cylinders and pistons and fuel injection and air-fuel ratios and timing and radio reception and electronic motors and A/C compressors and alternators and power brakes and traction control and ...

Just needs to know the interfaces to control and maintain the car.

We've made driving so simple that even a 16-year-old can do it.

The rise of modern graphics cards also empowers those small teams. Sure, you need a big team to create a heavily optimized scene if you want to win over the fps crowd, those who examine every shadow, but a small team with limited optimization can get 95% there simply because modern graphics cards are so powerful.

The graphics of The Forest (the second link by GP) remind me a lot of Crysis, a now 13 year old game. What was cutting edge 13 years ago is now commoditized because of vastly greater computing power and much better tooling (today's Unity editor is much better than what was available in 2007)

Or Subnautica: Below Zero, also made with unity. I'd hate to call that developer small, but compared to true AAA titles having less than 200 employees makes you small.

For a truly small games developer leveraging modern technology, look at Wube, the studio behind Factorio. Or Intoversion, the creators of Prison Architect. I think they each still have less than 20 people. Those games are not graphics beasts but they do rely heavily on modern CPU/memory speeds.

Doom and Quake were once upon a time AAA titles, and they surely were under 200 employees.

I just played Crisis, the original one. Still looks amazing.

The games you're showing have nothing special, it's just good lighting / textures off market. There is nothing complicated to display, almost no entitites, very small maps etc ...

Not wrong, but does it negate the quality of the games? Were they fun?

More telling, could they have sold better with a massive marketing campaign?

We're talking about graphics though, and those games are nothing special graphics wise. They look like something released 5-10 years ago, definitely not top end. This is not to slight the developers, they look good! It's always easier to make something of of the quality of 5-10 years ago. I disagree with the grandest-parent... top-tier will always be hard, for the foreseeable future.

Hmm. I don't want to find myself defending that it is easy. That is not my belief, either.

I don't think a lot of the complexity that goes into graphics is essential, though.

I do think top end will remain top end. However, I also think many games can be done without putting graphics rendering at the top of the budget. Axiom Verge is a recent find of mine that I feel hits this well. Thimbleweed Park is another good one here.

Which is not to say that either of those were easy. I doubt they were.

Reply on news.ycombinator.com