Building Worlds in No Man's Sky Using Math(s)
ฝัง
- เผยแพร่เมื่อ 16 เม.ย. 2017
- No Man's Sky is a science fiction game set in a near infinite procedurally generated universe. In this 2017 GDC talk, Hello Games' Sean Murray describes some of the most important technologies and interesting challenges behind generating both realistic and alien terrains without artistic input, using mathematics.
GDC talks cover a range of developmental topics including game design, programming, audio, visual arts, business management, production, online games, and much more. We post a fresh GDC video every weekday. Subscribe to the channel to stay on top of regular updates, and check out GDC Vault for thousands of more in-depth talks from our archives.
Follow us on Twitter
/ official_gdc
Check out our Facebook page for GDC exclusives
/ gamedevelopersconference
Visit our site: ubm.io/2ctNvqZ
![ชวนน้องลงทุ่ง - เงาะน้อยเพชรบ้านแพง X น้องมายด์ [ Music Video ] Official Saman Music](http://i.ytimg.com/vi/Bq7Hw9culjE/mqdefault.jpg)
"Y is no longer up" - as one of the main programmers for The Universim, which features fully spherical planets, this brought up an existential pain I thought I had left behind...
Never thought about that problem
Just by pure curiosity could you put a 2 point axis to solve this ?
i use a relative frame of reference anyway..
They have one of the best come back story in the industry. Respect to their dedication towards the game. Still giving out updates to this day. Bravo.
Waow, that was such a nice talk, i don't get why there are so many downvotes.
This guy is very passionate about what he is doing, please watch until the end its worth it.
Because he unintentionally lied a lot
@@Eternal_23 Na, there was a bunch of downvotes because the Gamer community is toxic as fuck and people are salty about No Mans Sky. They push the boundaries and because of it were a bit behind at launch. The result was a fan base attacking them which is ironic because those same people will cry they want something new but they want it with zero effort or growing pains. As I currently understand it No mans sky turned into a pretty decent game now, it just had some hitches at first and because of that you see nothing but down votes with anything that has its name in it.
No Man’s Sky needed some more time in the oven, and people were very unhappy about it being raw
King Nullpointer people unhappy because they've shown what they didn't had
@@kingnull2697 Idk what it was like then but I recently played it and I think it is awesome, There are a couple thinking I dont like, mainly the main antagonist, those little flying robots are annoying as fuck and I dont like the exhaustion meter and wish I could just run but the amount of content and things you can do in that game are down right amazing, It is one of the most amazing games I have seen in a while.
Fun to come back to this after the engoodening.
Would be fun to hear him give a second talk on where they've gone since this talk.
"the engoodening" 😂
Didnt expect him to show up in public without level 3 armor
dxxPacmanxxb Meh, without the safety of their parents' house and their keyboard, most Internet trolls are harmless ;-)
+James Grimwood Where do you live? Maybe we should visit your parent's home in which you reside, so you meet real life.
Lol case in point
BEST.COMMENT.EVER.
No Man's Sky sucked ass as a game but technically it was fantastic so gee, no wonder he shows up on a developer summit.
the talk begins at 17:56
Junpei Iori Cheeky
@@Luke-tu1yc someone's mad about their pre-order
I've watched the whole talk. I've always tried to imagine the technology behind the game. It always seemed really difficult to me. To build a game this big, this ambitious.
Now that I also know that it was such an small team, and that I know all the new things they tried when making the game, I think these guys did something not difficult, but impossible. And I admire them for that.
Say what you want about the game. If you are really interested in the technical details, and you understand a bit of them, you'll know that these guys just did something amazing, and maybe even changed the history of games forever.
This is actually not admirable. They did something incredibly dumb that wasted tens of thousands of hours in player time to make a...game??? I guess? Not really. Sean Murray looked up some math and created geometry that passed as a world and sold it for way more than it was worth. Should have gone into academia, but I guess he saw more money in this field. Regardless, this man should go down as one of the worst game designers in history.
@@notexactlypaul You should watch the video. He didn't spent years developing this game, apparently. No, he looked up the maths, made it and that was it. If you think game development is that easy then you're very mistaken. Also, this comment is 2 years old, why did you even bother replying?
@@catbreadloaf4416 In social sciences, often the basic work of the research is accomplished very quickly. The research often is taking existing data and finding a novel way to apply estimation to it. With modern computers, this task can be accomplished and tested in a few weeks on pretty much any laptop. Publishing that initial thought often takes a lot more work than that; you might gather more data to give your model more credibility, spend a lot of time writing, and have to go through a long peer review process. When it all works out, you have done a lot of work to support an initial idea that didn't take that long to develop. Sometimes those ideas are bad. You can either give up on the idea once your realize it or you can never realize it, peer reviewers never realize it, get it published and cited until someone figures out it was a bad idea. No Man's Sky is an example of a very bad idea being supported by years of work. The work doesn't fix the bad idea.
@@notexactlypaul 'Bad idea' is subjective. What makes it a bad idea?
@@notexactlypaul Yeah that's an opinion I strongly disagree with. In my opinion, he should go down as one of the best game designers in history.
"it was like building a rocket ship heading straight for the sun" hell yeah it was lol
And no one said "should we head for the sun, or that ice cream truck with lots of really nice people..." "Shut up, to the SUN!"
maximilianmus stans are the fucking worst people on earth
This was an absolutely enlightening talk, and since I'm working on my procGen terrain game right now, i'll be trying some things (like Domain Warping) ASAP.
Probably not one of the most beloved games, but definitely a landmark for procedural world generation, this is a gigantic step towards the future of game worlds, if we think of No Man's Sky more as an experiment, or playground, or simulation, than as a fully developed and conceptualized gaming experience, you learn to appreciate it a lot more
And the variation within a single planet i also pretty stunning. You can come up to nice plains or beautiful rolling hills with practical and exploitable terrain and on the same planet eventually find endless fields of insane cliffs and mountains where just finding a spot to land is near impossible for hundreds of miles.
A friend of mine sent me a screen shot of their planet and it looked amazing so I went out to check the planet and when I got there, the part I was flying over looked like a dead wasteland of arid cliffsides for a long while until I reached him and all the greenery and fauna started flourishing.
After around 16:00 when he's done talking about no mans sky and how much it got played, the talk got really interesting
That would have saved me 16 minutes if would have spent 10 seconds reading the comments.
Thank you. Skipped to 16, the prick is still insufferable but there's more info from this point on.
This will save me 16 minutes that I may never get back! :-D
Thank You dearly.
Lol, I just stopped working, started scrolling to the comments, and saw this one @~15:90. I was just about to go to my next video in the queue. I myself woul've stopped playing this game around about hour 4 or 5. But, since I paid $60 bucks for that thing, I played it all the way to the "end" in order to feel like I didn't waste my money. I would mute the audio, load up a podcast, and play through the game. I wonder how many people represented in his bar graph had similar experiences.
Thanks! I have no interest in the garbage that is no mans sky, but I am interested in what he has to say about procedural generation.
This has given me more information and use than all of my research on the subject put together. I've not spent too long but enough to say this video is outstanding. Thank you so much
Hello games is a company that I fully respect. They are an ambitious bunch of brilliant bastards. Sure, they slipped up a bit, but darn if they hadn't worked hard at it ever since. Love you guys.
it's a 20M$ scam. they haven't fixed the game one bit since. if it happened today they would've ended up behind the bars for sure.
@@neutralnarration1463 I get being angry, heck I get accusing them for scamming people. But to say that they haven't worked very hard to fix the game is downright factually wrong.
@@voldlifilm I _think_ what happened was, they had most of the features tested out in demo, but someone found a game-breaking error that stems all the way from the game engine or somewhere along the lines, causing them to paddle back to a really early build. Just a speculation, but that's my head-canon.
@@neutralnarration1463 you're drunk. The game is amazing.
as an audio developer this is really interesting to me, because the way he talks about noise generation shows that this is in fact a similiar way of thinking as realtime dsp algorithms. there you also can't just bake assetts into lookup tables before and have to manipulate the input a lot to get something out of it. and we work with noise functions a lot as well, even if not for terrain generation. wonder what would happen if some really hardcore dsp programmers started working for hello games and added some incredibly complex distortion to the uber noise function
Everytime they release an update, the game jumps to the top of steam sales. I actually kept my copy because I figured that they would keep adding more and more to NMS.
So far Im liking what they have done. Some things that I hope for in the future would be:
- More Interactivity with the animals (ahem... ark)
- Higher chances of fauna
- New sentinel types
- more animal types (especially for underwater and air)
- freighter vs freighter battles
with lots of weapons/upgrades for freighter.
- Some form of multiplayer (of course)
- Fixes to the aim sensitivity on console.
- Entirely new Multitool models
- More Story (I hear its in the works)
- New Alien races (because 3 alien types across 18 quintillion planets makes no sense. Make them procedural.)
- Ground based war between alien factions (Populating the planets with war, with various dangerous animals interacting would be cool)
like what if you could tame a nice TRex and mount it to attack aliens on a planet. Or have a few smaller dog type animals protect you as you walk around.
Another thing I found really interesting would be some sort of way to bring your favorite animal discoveries to your base. Like having a zoo with all of the craziest beats that you have hunted, except live ones would be chilling at yo house.
Just throwing ideas out there.
this guy is a fucking genius
NMS is quite a marketing fail but the information given in this talk is priceless!
2022 non dev (but still in the know).
Great expansion on the topic Sean, lots o great information for the pre-gen worlds for everyone. Love seeing how great your dream has progressed to be so awesome 😎
Ibam here after endurance update, I am doing some research about PGS for a project. This game has come a long way. Well done Hellogames.
Me with my little 2d voronoi Algorithm: "Wow, this was realy hard"
This guy:
this guy makes me wanna just go and make a game and enjoy it as much as I can
I am waiting for the day, when the No Man's Sky merging with The Sims, and the genetics works like in the Artbreeder. That would be the Ultimate mindblowing for me
Sims is for little kids, grow up
come on, didn't you get the point?@@sanjaymehta9253
@@sanjaymehta9253 probably you should grow up and let people enjoy what they want. Don't act like edgy teens.
Very interesting talk. Thank you all very mutch.
Amazing talk! Does anyone know if his code examples are posted somewhere, like github?
"You walk for ages and nothing happen"
OH REALLY
It's too perfect that he said that.
Just like in this one game...
I like NMS. played 130+ hours and I'm confident that it will get even cooler with every update.
yes it did not live up to every ones expectations. even not to mine, although I had very realistic expectations. but the guys at hello games are only humans and don't deserve hate.
the tech bedind the game is astonishing.
this was made by ~20 people.
This talk is evidence that they most definitely do deserve hate. Doing the math to proc gen a universe doesn't make a playable game, they should have known it doesn't, and anyone that has ever made games would have told them it doesn't. They did a fucked up thing and should be flogged for it.
@@notexactlypaul so you're obviously not paying attention to what the game has become.
@@notexactlypaul "oh no I don't know maths and everything with math in it is a hellspawn"
These GDC talks by Hello Games staff give a real insight into the technical marvel and artistic triumph that is NMS.
Perlin noise, also known as “how Bethesda creates their “worlds””
Such a great talk
and so today I passed 40 hr on playing NMS and this game keeps giving me breathtaking moments every time I land on planets.. I keep finding something new in each system. And btw I love the designs of Solar Starship, it's giving me a huge Treasure Planets and Starwars vibes
This started off boring as hell but then he seemed to relax and got into the details, maybe not deep enough but was interesting knowing what problems the team had and how they developed solutions. Was good to see Murray in his natural habitat the guy sucks at marketing but should definitely stick to more dev related issues. Would like to see more, was an interesting talk. Thanks for coming out of hiding :D
Really enjoyed this, and appreciate how open Sean was when talking about the noise generation code.
Can't get the thing behind float lfCurrentGain = lfGain + lfAmplifuFeatures; I guess this lf vars comes from another noise funcion but what's the deal setting de gain (commonly 0.5 in "plain" fbm) above or bellow for the entire x, y noise. I thought some octave driven value, but since it's at same level (ubernoise input parameters) as octave count, can't understand how this "Amplify Features" can give emphasis at some octaves...
Actually incredible talk even if the "maths" are a little beyond me. Definitely understand why those world generators were always a bit "off". Here's hoping those world generating programs are paying attention.
Seeing him after all the controversy is almost surreal
"Noise in the real world is exponentially distributed [ref: fields of noise for terrain generation]"
This game's complexity is underappreciated. I think it is incredible what they have done!
This is the reason the game improved soo much after that begining... He loves what he is doing!
You can say whatever you want about Sean Murray and the errors that were made with No Man's Sky, but is very obvious that he has passion in what he makes and was overwhelmed by the hype that was built around this game.
They tried their best to deliver the best product they could with what they had and i trully believe that Hello Games made an incredible job with this game, not saying that the flaws aren't true, but hating a video about the incredible technology that this games uses simply because Sean Murray is in it seems ridiculous to me.
Sean Murray outright lied about things that were in the game.
No he didn't it's evident that with No mans sky next, all these features he listed where ripped out the game just before launch as they where clearly not finished.
This aged well
@@stuwustudio These type of comments are trolls.
more comments about bad comments than bad comments.
so meta
elKONTEXO Now we just need to start comments about how many comments about bad comments there are to really confuse anyone looking at the comments section.
Hello Games is one of the game companies I strive to work with. I mean, they put their heard into game development.
you can hate this guy all you want, but goddamn he's a genius
Thanks Sean and GDC, this was interesting indeed!
The biggest takeaway from No Mans Sky marketing was, don't do it yourself. Imagine different marketing with it being "the biggest small game ever" and you can easily see how everyone wold have freaked out about how fun and great it is, even though it lacks content. Like Astroneer or Battlegrounds.
Anyway, I really liked the talk and insights. Super intereting even though I'm not a developer. Keep it up.
It's interesting to see peoples comments before the nms rehabilitation
let's put aside all we already talked about and solved to death.
let's enjoy and make use of the fact that for the first time, we can really hear and talk about all the technical stuff that interested and fascinated us from the start, and still does, even despite its shortcomings.
Where and how were those past mistakes and transgression ever "solved?" Somebody make a time machine, go back, and fix the damage and lies that game caused while I wasn't looking? No? Then nothing was fucking "solved."
i agree, I'm not a native speaker so I used the wrong word there. What I meant was the "things we/everyone already talked about to death". in my native language, i can say "we've been solving", meaning "we've been talking about in a serious way", so I mistakenly said it that way in english.
MidnightSt think about all those middleclass children that spent their hard earned allowance doing dishes for several days, only to find out that the computer game they spent hours on after school had major flaws. Think about how hard it must have been for these children, playing this on a computer made by other children in the third world, only to find it boring and wasting so much time playing video games with their friends...
Marius VanDamme =D. thinking about them so hard =D brings so much joy =D thank you =D
Their procedural tech is actually quite interesting.
Seeing Sean in public, talking about the game... It brings back all the hype... Good that some of the games magic is finally back again.
fair play to the guy, not many would stand up in front of tech savvy peers and admit he got stuff wrong, i wish him well in his future endeavours
okay they way he is doing the planets moving around is nuts and i dont get why he did it. same with marker moving. i could never hit that way-point so many times, just flying infinity towards a way point with planets flying around like wild.
People should get off of his back. He did work on Black and Burnout Series, and those were great. I was there (this talk) people joked there too, but everyone gave his team credit for trying.
Credit for trying but not for lying.
Awesome
awesome talk. thanks so much!
So... This is actually a little bit more complicated than I thought NMS's algorithm is.
It's almost, well, "you have to do serial things in parallel!" XD
The Man. Myth. Legend.
Didn't care for NMS, but this is very informative and super interesting lecture. It is good content that should not be voted down because it is somewhat attached to something that you did not like.
36:00 I see a jumping dear D:
He goes on about how many people were playing it, but check Steam Charts and see the massive drop off in interest after just a couple of weeks: steamcharts.com/app/275850 ... still interesting about the noise stuff, I hope it means more amazing games coming out with huge environments.
always great to listen to talks from passionate people. I'm not a programmer but this was very interesting
No man's sky Atlas Rises is one of the most interesting games I've played in decades. If these guys had the money and the resources they need, we would see an industry shift the likes of which doom made.
'Happy that Elite and Star Citizen are doing this now' ? They both were doing it before NMS.
No, they werent.
Awesome vid and sean is awesome too
Players who shat on No Man's Sky and got it refunded should watch this video and try to understand it
4gig banner art ... definitely game designers. if you're working on a psd over 1gig, you're asking for trouble. control them layers
game designers don't make banner art, illustrators do
Is Sean responsible for the girrafe glitch in BF3 🤣
And you finally get to that building.. and it's the same as the rest.
This highlights perfectly the problem Hello Games had; the problems they had were impossible to communicate to to the general public. Procedural generation of waypoints on non-baked assets using tangent on a point that doesn't exist yet?! They were doomed to begin with. I really feel for them as a programmer, yes he made some poor PR choices but he's a systems programmer with a small ass team. Not to get into the whole argument again but just wanted to see what other actual game devs think after seeing this talk
I play two games at the moment: No Man's Sky on my own and COD MW/Warzone when my mate are online. I can now see the irony that one game takes up 2GB and the other 250GB. As far as I'm concerned, the packaging that Hello Games managed is fantastic and, to me, THAT is proper video games design. Throw in the fact that any update that they've provided has improved the game immeasurably yet we're still struggling with DMR-gate on COD (even after a 'nerf') says it all. Props to them for sticking to it and continuing to deliver.
a very good talk, worth the watch.
ive enjoyed all the other talk[s] on no mans sky. they are good, technical & show a side people would not know, because they get on that bandwagon of the bad release & experience people were having.
its a shame because they are good talks.
but happy the game is still being worked on & updated. the tech & idea is fantastic & would of been a shame if it just went away, but this shows that sean is still wanting to carry on even despite the hate he has got.
17:30
"[we were everything]... marketing... not that we're very good with that kind of thing"
This man *can* tell the truth!
3-4 years!!!!!?
Give that man props for sharing his tech instead of focusing on how you hate the product that he made with it.
Holy shit he's on camera!
Using 6 month intervals, i got an average of 7 team members.
The problem is that they used ONE method for ALL planets...even if the world generation is quite good, you will see the same shapes on every planet. It would be way more flexible, if you would have different generator arrays for different types of planets...because...just because !
this was a really good talk... especially the procedural bitangential waypoints one...
Such a great talk. Now I am going to play NMS, as this is the game made by truly passionate people thus it is doomed to be good.
He starts talking about world gen at 17:55
wtf hes alive!
This game covered a lot of ground in the procedural generation area. A lot of game got inspired by this, even if no man's sky didn't turn that great they kind of lead the way. I've been playing a lot of proceduraly generated games and I am well aware of the limits of procedural content generation but I've always saw the potential in this and I was really happy when I saw what they did at hello games I'm still really happy to see that no man's sky pushed back these limits and show to the developer community the benefits and potential in procedural generation done right (the the gameplay or artistic choice here, but the real-time content genration).
anyway, good talk! thumbs up and don't forget, blame Sony and their marketing team not the developers, they are trying their best !
What games, exactly?
I'd say that the way Elite Dangerous and Star Citizens implemented there generation solution was inspired by the work of Hello Games, the solution is far more complete than what we saw in, let say, Mass Effects for example. Astroneer was likely inspired by both minecraft and No man's sky and I see a couple of indie games that try to build on the idea of no man's sky while avoiding the caveat.
Josué Boisvert Elite dangerous came out in 2014. No Man's Sky was released in 2016.
Oh yeah, that's right. well, cross that one off sorry! But my point was that the game as a fantastic procedural generation system and it really show's off the potential of this methode. Even if the game is far from perfect, it inspire to use this kind of solution to build even more amazing games!
I'd say, if anything, it showcases the horrors and pitfalls of procedural generation to a glaring degree. The game is worse than Spore and produces far less interesting fauna. It is hardly the first game to have procedural generation but most games handle it responsibly, No Man's Sky not so much and it has served as a cautionary tale of failure more than an inspiration.
WOOHOO, SEAN MURRAY
The technology behind No Man's Sky has incredible potential for games, and to ignore it because of the game it's attached to would be childish and stupid.
cavemancolton wasnt hugely innovational though... been done before, but i agree, massive project
cavemancolton don't talk nonsense
Danny VD hm i think the expectations were a little high from certain lies to the media too...
I think saying its a good game or not is pretty subjective and opinion based and thus not really suiting a GDC video, where learning mechanics and doing something productive is key, so complaining about the haters is benign and just makes the haters hate more, which makes a bigger deal than it already is
Danny VD ohh if you think about it as an interactive screen saver and not a game, then its great! Too bad is sold as a 60 bucks video game because its not worth more than 5.
No Mans Sky was a beautiful space photography simulator.
needed more Pokemon!
honestly, when I hear no man's sky I just hear wasted potential
Y'all should check it today
@@vaisakhchandranb9643 Yeah, you're right. It'd have been better if Ubisoft did it
smh
I wonder what the name of the engines he made are.
He didn't make any engine(s). He contributed to one or more. At that level, no one person *makes* an engine.
I still find new stuff and I got over a hundred hours into the game.
I understand very little about this but I just love hearing Sean Murray talk about this shit.
Frack, I thought I was the first person to re-interpret RL map data for game content generation. :-(
The video was published a minute ago and already there is two dislikes.
People are children.
I know multiple people that aren't children. In fact the majority of people I know aren't children.
only children resent being scammed?
+ItsAllGoodGames
No, but children have trouble separating a talk about the tech behind a game, and the game itself.
+Alexander Abdulov You're right, there should be like 420 dislikes in the beginning. Hail Sean Murray!
It's a fantastic talk about the pattern he got to and it turns out no mans sky is quite a engineering marvel when he mentioned the game size break down. Having said that it really didn't warrant the price tag and as ignorant as that sounds together I think it makes sense- he pretty much slipped and said what the game was in an off hand comment- it's a fancy terrain generator.
If there is a sequel, or just a next project for Hello Games, whether they build upon NMS or something completely different, I think it will be successful.
He worked on Burnout 3? The single greatest racing game in all of history??!? Respect.
OUCH and OOF
This shows how great technology doesn't automatically make a good game.
This was a great talk but at around 11:18 did he just pronounced heuristics as "horistics"
Sean is the man! What that small team accomplished is incredible. The tech is undeniable. The app was just once use thereof. The potential is staggering.
17:25 He didn't lie about the marketing thumbs up
Personal/commercial projects aside, if what he's doing helps the industry then that's great.
I think he's a very smart guy. It was just a bad decision on Sony's part to let him do the marketing for the game...
Vlad Mihai shut up
Say what you want about ma man Murray, but the tech in this game is impressive and he seems to know he's way around mathS