Video není dostupné.
Omlouváme se.
Why Do We Need Polygons In 3D Video Games?
Vložit
- čas přidán 19. 11. 2017
- What are polygons, and how do they make our favorite games three-dimensional? Let's explain.
Subscribe for more: / gameranxtv
Footage via: / @bel_itman
Video Credit [1:10 -1:22]: Bart de Vries
• "Autumn Valley" - Styl...
(www.artstation...)
im a game artist myself and I have to say thing was pretty spot on. The idea of moving to something beyond polygons has never even occurred to me simply because of how well the system works. That being said the way we use polygons is rapidly changing. Before everything was done by hand. start with a cube, add extra edges, and morph it into your desired shape. Now most studios sculpt their assets using millions of polygons only to drastically low the amount near the final process. We are now seeing a new method in the industry. People are starting to use photo scanned object to create "perfect" models. We're quite a ways off but looks very promising. With that in mind this type of technology does nothing for stylized art, so it will never fully take over. Its amazing to see how far we've come, looking forward to what the future holds.
i do 3d modeling too, doing it in collage next year :D
Tiedie true, I am making a game using only my own assets, and I am manually shaping cubes/basic shapes in blender into the shape I want, but I am no artist and so I have gone for a low poly design since I just don't have the time/patience/skill to create high poly models.
Yo 3D artist here too.
Its interesting to see you here Tiedie I've been subscribed to you for ages now how have a career in game design one day
"A pixel is a dot"
*shows square*
Dpi
A pixel is technically not even a square depending on the screen.
A dot is just how something is appearing to the eye,in a nutshell.
A dot is actually a square. Not a dot.
If you are watching it on Trinitron where each pixel are zig-zag itself, it makes you think, is it a pixel, or square, or dot, or zig-zag RGB, or maybe triangle...
As a graphics programmer myself, videos like this usually make me cringe hard, because they're (usually) ~2/3 flat out false. Like putting technical terms and basic words to form a sentence into a blender and see what happens. Luckily, this video is not one of those.
Well done!
That we're not doing full blown molecular simulation, but only surfaces (like what the brain and eyes are doing anyway) is the whole point of realtime CG. Calling it a "magic show" is a great paraphrase.
3d designer here. Great video, one thing I want to add, in the future when the hardware is strong enough we might see a shift from polygonal modeling to point cloud based object scanning and rendering. Point cloud is like a cloud of "atoms" that are represented with points and the software connects all these points into a model which is much more detailed then polygonal model
Falcon is the kind of guy that mixes his tea with coffee, because "He likes trying new things out"
The Bakersman having done this by accident I do not recommend.
Doug demuro fan spotted.
The Bakersman I mix Pepsi and Coca Cola with Caffeinated Coffee, works wonders. Just don't replace Pepsi and Coca Cola with Red Bull...
I've seen people put creamer in coca cola. This doesnt sound too bad.
MegaNocab i agree😂
I've been enjoying these technical videos that Falcon does every so often. Even if they are simplified explanations, I learn something new.
same
bbarrett726 k
bbarrett726 agreed I want more of these
The more you know
Right falcon got the best topics on this channel
UE5 running 16 billion polygon on a PS5 how far we've come.
Actually, UE5 on PS5 don't have limitation on Polygon anymore. They explain all polygon are not limited. So we are in the end games is all about IO now and SSD speed to stream all assets without limited by polygon.
@@mightyhadi6132 For Nanite, yes, but Nanite has the downside of requiring heavy precomputation. It doesn't work so well for deforming meshes (ex: character animations) and it's completely unusable for dynamic topology changes (ex: fracturing an asteroid when you shoot it at precisely the point you shot it).
@@darkengine5931 agree , that's why the new UE 5 2.0 are enabling Tessellation , for deformation . Fortunately PS5 is running at higher clock speed and has high fixed operation .
@@mightyhadi6132 Dynamic tessellation is awesome although I think hypothetically we could do even better with per-pixel displacement with a raytracer... but I think real-time raytracing isn't quite there yet.
We don't have a single game that look close to UE5 demo's lol😂
I'm a 3D artist and this is spot on and I actually don't think there will ever be a need to change the polygon system, I mean sure in the near future as computers become more powerful we wouldn't need to tweak 3D models as much and they can have very high polygon counts but we probably will still use polygons nonetheless because they do the job and they do it smoothly and flawlessly if enough knowledge is applied.
I came for polygons...left with an existential crisis.
Khalid Himmo Nah, the actual world is a bit stranger. Wave like voxels moving through 26 dimensions.
ron
Like ur mom
Scotty Nguyen's 80s Mullet then u r his brother?
jajajaja
Khalid Himmo You should watch the video "Are You a Simulation?" on VSauce3.
+ronindebeatrice Games need more voxels.
*"...We 've created an art form that combines so many different technical, mathematical, scientific methods of processing information, to create something that suspends our disbelief, and allows us to engage in another world."*
'nuff said
Beautifully said.
He literally just said that as I read this.... Woah.
@@CloroxBleach-oj2rl Carl Jung's Synchronicity theory.
"we play video games not because they look perfect but because we enjoy em"
rtx : hi
people : OHMYGODREFLECTIONSTAKEMYMONEY
Clen Vingt-trois it depends on the game. realistic games should look realistic. cartoony games not so much. both still benefit from good lighting.
I went to school for digital animation and made a few apps using some things I learned. I ended up pursuing other things because my heart was just not in working for a big company, or wrestling with software and rendering for days on end... and I was more in love with hand drawn animation to boot.
As far as tech goes, the new thing will be photogrammetry, and have done some myself for fun. I love it, and like others have said, artistic expression will always have an appeal to audiences. The most mind boggling thing to me are the people who were smart enough to write the programs and the algorithms to be used as tools by the industry. Writing code so to have a computer replicate light refractions and render it realistically inside of a glass surface cannot be easy.
triangle are also used because they are the only polygon which is monoplanar, meaning no matter how you arrange it's vertexes in 3d space it always occupies only a single plane.
Exactly! Those f*ckers can´t be bent, therefore it avoids a whole new level of glitches caused by any other more complex face.
Zenthex TRIIIIIIIIIIIIIIIS are annyoinf to work with
Normally you can work with different shapes in 3D Editors. Then it exports them as triangles, but you are not limited to triangles when working, for example, in 3D Studio Max.
you can use quads for relatively flat shapes for example
Yes sir!
Theoretically, if you had computers that are powerful enough, you could create a virtual table of elements: tiny hexagonal polygons like atoms which are meant to stack on top of each other based on a certain set of rules similar to physics. You wouldn't have to make it the same "resolution" as real life atoms. Pretty soon it might actually be possible to render 3D objects based on a virtual table of elements. You could simulate chemical reactions based in virtual time and predict how those chemicals might react with other chemicals. What this means for video games is not that important, but for chemists this would be an immensely powerful tool. You would no longer need to worry about time in your experiments. Simply input the chemical parameters into the simulation and press the go button. 100000 simulated reactions occur in a few seconds. For the medical industry this would be incredible. Nanobots would become the next big thing.
Merry Christmas
Whenever i see these kind of Gameranx videos, i can't help but think that at some point near the end, Falcon's voice will increase gradually as well with the music in an epic-like way. finishing with a message of faith and hope to humanity... or some shit like that lol.
Awesome video! =D I love these random video explaining stuff... even though I pretty much already knew most of it, it was fun to watch.
The circle is made of squares
The circle is made of squares
The circle is made of squares
The circle is made of squares
**Triangles**
Yep, triangles. If you're doing it right, it only takes 2 triangles; you'd use a shader to render the curvature.
lol 😂
i can just imagine you absalutely paranoid*Circles are made with sqaures...*
Let my elevator alone
Let my elevator alone
Let my elevator alone
Let my elevator alone
Simon Nitzsche Pollygons are rectangles sometimes. height maps and normal maps help draw circles on flat surfaces to : p
polygons...... *I LIKE POLYGONS*
chizzy but i hate math
I like porygons more
*I LIKE TURTLES*
Why do i see you everywhere?
The more polygons on a 3D game the better it looks. For example, Heavy Rain video game.
Polygons are cool but atoms are where its at.
Raymarching can render perfect 3d object
I always wanted to make video games ever since I was a kid, and now I'm actually starting to make games with unreal engine 4 and cryengine, so this video helped me understand the science of make a video game. Thank you guys so much!
Who thought from 2:02 this game was GTA San Andreas graphics before completely rendering to GTA V graphics?
Is it just me or did I hear him say,"Videogams"?
No’ît koxöza.
He says it several times in the video. He even says "sam" at one point instead of "same". I have heard many different accents before but this is new to me.
Lol. He said that.
2:07 That's basically San Andreas
You do know that's GTA V, and GTA V is based in San Andreas just like GTA San Andreas.
Ron Da Bird you didn't know what I meant. I was talking about the graphics around 2:20
He was saying that the graphics at that point in the video looked similar to the graphics in GTA San Andreas.
Which video ?
nop, san andreas was lower poly count and got worst texture ( except if you use mod ;P)
and thatboi hops it's wrong, Gta V is based in southern San Andreas, 2/3 of san andreas map is missing ( San Fierro county and LV)
Very soon
"Why do we need colors in video games?"
"Why do we need sounds in video games?"
"Why do we need stereo and more in video games?"
"Why do we need keyboard and mouse in video games?"
"Why do we need gamepad (for noobs on PC) in video games?"
"Why do we need no more EA in video games?"
etc.
:D
Whas it even actually explained why do we need those polygons? Have I missed it? He didin't even shown any alternative...
Why do gamers nowadays cares more about game graphics, rather than gameplay?
Because... there is ni gameplay nowadays?
He did mention an alternative but that alternative is too expensive processing wise, which is rendering 3D models like a voxel image, with algorithms defining a shape.
Jay Co
I'm pretty sure that even voxels use polygons. Though I have no idea how you'd create a perfect sphere using a 3D application, but I suppose you'd do something similar to how you create a perfect circle in 2D. That would require an insane number of polygons though, so how does one create a 3D sphere, in a 3D environment, without polygons?
polygon is the best pokemon
I dunno if it's the best but it sure is cute, also i don't remember seeing him that much in the anime..
the one from gen 1 is pretty cool
+JormungandR - Porygon was never banned, *The* porygon episode was.
He's appeared in other episodes of the show, but it's never been a popular Pokemon. It's like saying "Wow.. I don't remember seeing Belsprout in Pokemon.. must have been banned."
What would it take for a game to have a fully rendered and realized environment? Where you use a shovel and the dirt gets picked up seamlessly or you can take a bucket off water out of a bath and the water in the bath is lower. That will be amazing.
Tim Sweeney believes we'll have true photorealism with 40 Tflops of GPU power. So in roughly 7-8 years.
Realtime fluid physics at a large scale are further out as a standard in games, as it requires enormous amounts of processing. I wouldn't expect the oceans of an Assassin's Creed game, for example, to use realtime fluid physics that rivals real life for another 20 years.
Well technicaly it could be done if the only thing youre doing in the "game" was taking water out off bath
Right? lol!
Martin Boylan at that point you may as well just live in real life
But I can do all those weird things while in the comfort of my home, clean and spotless.
I am Indian game dev myself and I make my 3d models for my game art but I haven't think about why we use polygon before.
But now I think it's the best way to make a 3-dimensional game for sure
Keep in mind polygons aren't the only way to represent 3D models. Voxels are another famous example. There's even 2 DOS games from the 1990s that scaled and rotated ovals instead of polygons. They're the Ecstatica series, by the way.
3d modeling programs like blender use light beams to render images, say you were rendering something at a 100x100 resolution (i know this is very small) the program fires beams out of all those pixels that bounce off objects differently, it depends on what material you make it, then if the beam hits a light it will show up as that pixel on screen, in a complex scene in blender it will have to do this about 2500 times per pixel, thats why it takes sometimes weeks to render animations like this, and thats also why video games dont do this right now because they would proberly run at about 0.00001 fps.
That is more of a lighting and shadows question though.
Also with Blender, it doesn't do anything light related until you hit the switch on one of rendering engines it has, and even there:
Blender Internal does only 1 calculation for 1 pixel for X lamps you have in the scene, either if it's lit by it or not(unless you use smoothing then 15 samples is enough). To be honest, games would run with Blender Internal at like 5FPS.
Blender Game engine does stuff like a game engine and does more smoke and mirror stuff by predicting where the shadow would be.
Cycles is I believe the one you were talking about with that. A lot of ray scattering and... even more of ray scattering.
Eevee(coming in 2.8) is kinda like game engine, but it's a mix between raytracing and modern game engines, where it's 99% realtime, but if you look close enough you would still see it losing noise over time every time you stop moving the view around.
QuickscopingFTW Didn't Carmack have a raytracing engine running in game in near real time 20 years ago?
+ronindebeatrice
Yeah, but that was like 320x1 raytracing with just 1 simple thing - see which part of the wall/object/whatever's texture it needs to render for that pixel, and the distance. Reason that the height resolution is 1 is beause it does only 1 check for a whole vertical line on the screen, and then draws that whole line based on 3 variables - what texture to use, what part of the texture, and how close it is(how big to draw it).
It's maybe a bit more complicated than that(transparency, for example, drawing hand of an enemy and a wall at the same time), but still, same logic applies - non-realtime raytracing with a lot of calculations and Wolfenstein 3D's simple near-realtime raytracing aren't the same thing.
Architector #4 I wasn't thinking Wolfenstein actually, but something much later.
What is it then?
Can't think up of any other raytracing thing Carmack did long time ago...
@1:53 Video gam
I typed up a comment for this just now, looked to see if anyone else did already, and now I'm sad.
Better luck next tim
gta 5
MISSION FAILED! we'll get em next time
Alpha Morty I was wondering if anyone else has noticed this. In all these videos, most of the time when he says "game" on it's own it comes out as "Gam", but things like "video games" and "gameranx" always sound fine lmao
Damn, you just made me love this beautiful art form even more. Very well done video.
video gam 1:50
Vido gam
Off-topic, but I can't be the only one who has the urge to punch the upside-down sunglasses douche in the face.
That's just the movie Hangover
OSHMKUFAHSA He's talking about the way falcon says "video game".
vidyo gam
Great video, interesting stuff
Thanks!
gameranx Senpai has observed me.
*commenting intensities*
cocaine is better man
next question: why does EA still exist?
For sims
its in the game
To make shitty games like skate
99.99 for skateboard
Because they're still making enough money to stay in business. It's pretty simple.
More polygons means
1. More memory needed to remember those vertices
2. More processesing needed to calculate their position and manipulation happening to those vertices
3. At the end we end up buying high end graphic card with highest memory and also buying high end processor.
I'm currently taking a computer graphics class, and the breakdown of the matrix math used to render, transform, and shade even the simplest scene or object is absolutely mind-boggling. I'm thankful that we've had such incredibly smart people to make the field of computer graphics out of nothing, and build upon it to the point we're at today.
As a 3D artist, making models, making a decent model takes no less than 200k polygons. Video games are really behind in that aspect, just having around 15k each
15k is quite low for AAA games. Most models are double that, with certain key models being 3-4x that amount.
Blueberry Bandit pc or consoles?
Polycount isn't that heavy that you are limited to 15k per model console and Pc games have models with 200k plus polys especially vehicles. (Lead 3D artist here)
Diego Salazar Consoles, and specifically for characters. I've seen models with as high as 85k on console.
for something like a gun, I usually stay under 10k verts
What about that company, Euclideon, I️ think? They claimed that they found a way to get rid of polygons and use molecules.
years ago. There's a reason there's no news about that. There's no real-time application for it, as it's not actually feasible.
Bearded Weirdo TV Thank you. There is an indie game thats out that uses it.
DwayneW1987 Do you recall the name of the indie game?
yea I remember watching tht 6 years ago, but we dont hav the tech yet for everyone, the general consumer, to hav a system tht can play those kind of games. (plus its still too expensive.)
The company is still in good business, and they do use the technology. But they seem to have stopped following the plan to use it in games. www.euclideonholographics.com/
I guess at some point they came to some technical difficulties that they just can't overcome yet.
I don't know what's more complex and more artistic , the way the developer throws polygons to transform into video game art or the way the video narrator throws his words on this video and make it sounds like poetry , very enjoyable video
Isn't that's why we play videogames so we can do things we can't do in real life?
Like dating sims for example
I play them for research.
Gamer dude89 and if you enjoy doing those things witch you can’t do in real life then the main purpose is for enjoyment
Or used for training solders.
While we pretend they have no effect on people because it's "just a game".
*eats my cake*
I play gta v to go into the nightclub
I don’t know about anyone else but after watching this video it made me appreciate the work of video game developers and how far we’ve come (I remember thinking Ocarina of Time as the pinnacle of graphics) to be able to experience this..........I’m baked. Great video as always!
When I was younger, I thought games like GTA V were made PIXEL BY PIXEL.
I do a lot of modelling, not for games but for my own personal animations. This video was great and really explained polygons in a way that a lot of people can understand so thanks for that c: In a way, 3d modelling and animating is just as fun as playing games because you get to challenge yourself, improve and when you reach the end goal, you feel a great deal of satisfaction and you feel proud that you've created something that was originally just in your head. It's great!
I love it when falcon says a long word like philosophical or mathematical it’s just so,ENLIGHTENING
Oh yeah, Falcon!! I always cringe clicking a Gameranx vid in case that other guy is hosting.
why is this 3 year old comment with 2 likes the top comment
Knack: 200,000
Knack 2: 500,000
knack 1 and 2: still dog shit
OH OOOHH DID SOMEONE SAY KNAAAAACKK!?!?!?!
+Sam Rod You're uneducated Knack is the game of the year, every year.
@RandomFinn you mean super Mario bros 2 baby
Good vid mate! As a cg character and prop artist myself i have to say that the amount of polygons is not really the problem anymore. These days we use what is called a BPR render method. You sculpt a high poly character (lets say 20 million polygons) and then you create a low poly character ( its called retopology). After that we create bump maps ( like texture Maps, but with hight and detail information) based on the high poly model. we wrap the bump map around the low poly character and if you render your low poly character, it will look just like the high poly model with the same details and smoothness. Although we do use more polygons to create more detailed characters and environments, next to polygons the shaders and lighting is what makes the game look good and heavy to run.
It's MIND BLOWING how far video games have progressed and continue to do so visually.
☢️❎⚠️(Ive not watched the video all the way trough) BUT.....
whats with the idea from Euclideon? They made the atom system nearly perfect so that it could even run on a laptop.
You can download it right now its called Geoverse!
oh and sry for my english
Maybe in the future we will find new ways of creating graphics!
Amazing video btw ♥
Back in high school in 1997, I worked with amazing geniuses who developed a real time Spline engine. Splines are mathematical curves, giving the engine a formula to create models, rather than beancount polygons. Splines were already used in pre-rendered high quality cutscenes, but not in a game engine.
It worked beautifully, especially with scale, as a formula is more efficient than beancounting, as the complexity goes up.
To this day I don't understand why it wasn't replicated again. A spline can have customized detail, you can calculate more or less curvature, giving a huge range of performance and depth.
And here's to you guys, Tony and Chris, smartest guys I've ever known, who created that engine at age 15
I'm sure a bunch of you are wondering how we can make such " Low poly models" look so detailed in terms of shading and smoothness, how we make a crack in some bricks appear without actually modelling the crack.
This is something we call baking textures, and it does actually involve modelling small details and then using a technique called "Retopology" which is essentially creating a simplified silhouette of the detailed model.
Once we have that silhouette, we can bake the shading and small details of the original model onto the simplified one as a texture (Basically just a picture that wraps around a model perfectly), the computer then uses these pictures in different ways and calculates shading and details depending on the tones and colours of these pictures, and then simulates them in 3D space without actually creating any unnecessary polygons.
If you want to know more about this process, look up "Baking textures" on google, you may also want to look up retopology if you're wondering how the process of creating a low poly mesh out of a high poly mesh is accomplished.
A real question would be is there a such thing as an 10/10 game ?
thatoneguy from the internet SMG
thatoneguy from the internet 3 words:
Super
Mario
Odyssey
Raul Andres But would it still be if it has glitches in it ?
Sir Galahad true so the last of us is 9 out of10 in that world
Undertale. It got 10/10 on steam and IGN
I'm a simple man. I see a Falcon video, I watch it.
I've been a level designer and 3d coder for more than a decade now and I must say your explanation is among the best I've ever seen. Simple, down to the point, true and informative.
I very much like at the end when Falcon said, "It's a magic show." That sums it up for me. I used to produce animation, and even when we did 2D animation, it was all done in the computer, meaning, we were creating the illusion of traditional 2D animation with more advanced technology. Magic, sort of. Also, I'm not a particularly technical person, so I appreciate the simplified insight as well.
Games are becoming more and more graphically impressive, I wonder how long it'll take before the advances plateau. With that said VR will grow and grow over the next 20 years so it'll be exciting to see that really flourish as it has a long way to go at the moment.
You could argue that it may have already hit it's first plateau. Ultimately the goal is to create a piece of art that is pleasing to look at and conveys the message of the story. Something you can get lost in and forget you're not part of that world. I'd say they hit that level a generation or two ago. It isn't a coincidence that so many games are coming out with less realistic ascetics. It isn't just because it's cheaper (Though I'm sure that helps). It's because the artistic direction has long since become more important than our ability to make something look believable. Mostly because realistic is something we can already do at any level we'd need to.
There's a second plateau. One that strives for indistinguishable realism in games. That's an artistic goal all it's own. But if you ask me. Success like that is now just a goal for those pushing the limits for that wow factor. Not something every developer needs to, or even should, strive for. These days it's about what's good for the game. Not so much about what's going to show off the power of the system.
Now to undermine myself a bit. I just got done with Evil within 2. Who animated the faces in that game? A Barbie doll sculptor?
Frank Stapleton I never said whether they were good or bad but the technology is in its infancy, it has a long way to go before It hits what many see in sci fi movies
Hardly surprisingly considering that was released 25 years ago. I mean it's pretty obvious VR is still in it's infancy, the technology has only just begun and has a lot of restrictions still despite some fun games.
"Games are becoming more and more graphically impressive" - on screenshots, maybe (still, they are just copy-pasting "reality", not really impressive for alter/digital world), but in action they are becoming more and more annoying.
Soon all games will look same, they already do... As for me, it is quite boring and monotonous, and really not Impressive (Well, maybe, only if for a person with complexes or with psychological problems, who do not live, but simply exists)
Project Gamer If you're looking for a plateau, we've been in the 256 bit era for over two console generations. Even the Xbox One X and PS4 Pro are 256 bit consoles, as was the original Wii. Let that sink in for a moment. The only differences are processing speed, RAM, resolution, and a few other factors, but other than that we've essentially been using the same outdated tech for a while now.
Because polygamy is the way to go. The more the better, right?
this is super informational and entertaining to watch, I personally am not a tech savvy person but I am studying in graphic design degree and it's definitely interesting to learn atleast the simplified version of whats going behind the scene that we all love and enjoy in video games, hopefully I can finish my degree and contribute to a gaming company
As someone that fiddles with 3D thingies and puts pixels together for a hobby, I would never be able to go on about my passion as basicaly and at the same time inspiring as you guys.
As an amateur artist I never even considered the use of particles simulation to make "real things" other than fluids, grains or clusters of things, much less the possibility of using curves instead of faces to make actual meshes!
This was one of these times when I think I'm gonna hear more of the same then someone totally baffles me by showing how much I don't know yet, even basic things.
Kuddos for gameranx. The peeps that made me consider giving up on amateurism and the fear of math to learn what the computer actually does and fully commit to digital arts!
As a computer scientist I must tell you that a Polygon is not a 2D shape, it's actually abstract data. The rasterized frame (which is composed of pixels, not polygons) is indeed 2D but the polygon itself is an abstract form of data which describes how the computer should draw the 2D frames on the screen in a process known as "Rasterization". And it's a 2D picture that represents 3D things in a perspective view. In essence it's a perspective drawing done by a computer. BUT polygons can actually be displayed in REAL 3D space as well using LED cubes (RGB LEDs are used as pixels in 3 dimensions, Physical dimensions by the way). But yeah a polygon is fundamentally different from a pixel in a sense that a pixel is something physical while the polygon is abstract data. The Polygon can become physical if you have a proper 3D display that actually takes real space, such as a LED cube. The 3D screens that we have are actually false 3D, it's not actually taking space in 3 real dimensions. A LED cube actually does take place in 3 REAL freaking dimensions, I hope this is clear. Polygons are usually abstract but they can be materialized and become "physical" sort of speaking...
As a computer scientist... this is a very confused and also quite misleading explanation. A polygon is a 2D shape. A triangle in 3D space is a subset of a 2D plane. Mathematically, there is no question here. Of course geometric objects are represented using abstract data (as far as the electronics is concerned, it's just a blob of zeroes and ones), but semantically, they are geometric objects.
Triangles are continuous (they are literally a continuum), and pixels are discrete point samples. The process of turning polygons into pixels is sampling. Right now, we know how to sample polygons very efficiently in hardware. We don't how how to sample other kinds of curved shapes, or volumes. Or, rather, we do know how to do it, but exactly how you do it depends on the specific kind of shape or volume.
Polygons (and triangles in particular) are a "sweet spot": anything which can be understood in terms of 2D differential geometry can be approximated by triangles reasonably efficiently, and triangles can be sampled reasonably efficiently.
Well, higher order interpolation techniques exist, but let's just stick to the one with the lowest computational requirement, straight lines.
One key advantage of straight lines is that they are still straight lines if you apply a projective transformation (e.g. a perspective camera transformation) to them. Even other conic sections don't transform as neatly (e.g. circles can transform to ellipses or hyperbolae).
Well, object oriented programming can let you define properties of the vertexes so you can choose between different primitives. This happens when you need to do ray tracing kind of stuff.
That's not real-time game rendering.
Modern production ray tracers tend to tesselate curved surfaces into discrete differential geometry shells or micropolygon meshes (which are topologically the same thing). Directly ray tracing anything nontrivial, such as spline surfaces or subdivision surfaces, is too expensive and too numerically unstable, and besides, there's no other sensible way to implement displacement. Only simplistic cases, which tend not to occur in animation or vfx production, benefit from not using tesselation.
It's also very uncommon for modern production ray tracers to use OOP specifically to attach data to vertices or surfaces because it's too inefficient (in both RAM and cache usage). OOP tends to be reserved for higher-level structures.
Actually, the question is: How the heck would you create 3D computer graphics without polygons?
Technically you're dealing with pixels regardless of how you render objects. But I'm assuming, though not certain, that you just bypass the vertex shading and go straight to pixel shading (or whatever method of drawing the pixel that's used). From what I understand, you need to use ray-tracing or ray-casting. I know a little about ray-tracing now, so I have a bit of an idea how it's all done now, but I'm still not sure.
I'm guessing without polygons, you're using a center point and the width, depth, and height of a voxel (or just a width if you're using cubic voxels) to determine whether or not a ray (from camera position) strikes a specific voxel (through a pixel on the screen). Then you just draw the pixel based on the colour of the first voxel struck.
I'm going to look into it in more detail someday (hopefully soon).
You could use bezier curves, so that the mesh is infinitlely defined (kind of). check out this demo, not really pratical for rendering or modeling, but still show the point: czcams.com/video/KYekhnLHGms/video.html
Nacim DarkDrifter
In other words, use one or more mathematical calculations to draw each pixel individually like you might do if you wanted to draw a circle in two dimensions.
Nice question. I'd say make your 3D models with spheres, then cast a ray from each pixel on the screen to hit the sphere, and there it is. Another way is your models could be made of voxels each the size of a pixel, and do the raycast thing again to get the voxels nearest the screen. Another way is to make everything out of billboard points and draw an image to the screen at that point.
I've done 3D work for a while, and the short answer to the short answer is this: we don't need polygons, but they make everything we do in 3D much easier. Early 3D games such as Mario Kart and Wolfenstein 3D didn't use polygons. It was simple, but it worked for the hardware at the time. Once we had better computing technology, polygons were used for their flexible nature and that they could represent a 3D object in any conceivable condition. They define a surface and that surface could be acted upon to get a variety of visual effects. Change the brightness and color based on lighting. Change the look by applying materials. Change their position and orientation but retain the identifiable look.
Since the beginning, though, there have been attempts to move away from polygons or at least supplement them in 3D space. Quake 3 had curved surfaces that were truly curved using complex mathematical equations to define the surface, but polygons were still used in conjunction. There have been a few voxel (volumetric pixel) games that use a series of points in a 3D space to build objects, but they tend to stay simple and blocky. There's been interesting work done by Euclideon using massive amounts of points in space scanned from the real world, but it seems very limited in use as objects can't be lit or even moved. You can also create whole 3D models using NURBS, surfaces that are defined using math and specified points in space, but it's difficult to get a high level of detail using them and hardware isn't set up to render them.
For now, we're "stuck" using polygon objects for their quality, versatility, and ease of creation.
I remember not beeing interested in this video and now i just used it for help in a school project
nice
What's the game at 1:09 - 1:21??
Really like the artstyle
World of Warcraft
Why Do We Need Atoms To Exist?
"macintosh+ 420 starts playing"
Funny how a sophisticated structure of atoms is asking why it needs atoms to exist.
It isn't that we "need" them to per say, they just do. I won't say "simply" because there is nothing simple about them.
WooferJr that's why I believe in God cause of how amazing we are perfectly structured. :) either that or believe in random lucky chance that we exist only to question our existence.
Those atoms may be virtual and be nothing more than a simulation.
A point missed in the video is the use of trigonometry functions thousands of years old dating to 1 BC needed to transform 3d points (x,y,z) in space and "project" into 2D, which is what we all see on a screen (or papyrus back then). Modern processors and GPUs have these sine/cosine matrix projection formulas built-in to accelerate the process, which makes high-resolution 3d animation on a screen possible. If you're curious, the formula to rotate a point around the x axis is newx=x; newy=cos(theta)*y - sin(theta)*z; newz=sin(theta)*y+cos(theta)*z, where theta is the desired angle.
"What makes videogames magical is not that they are perfect, is that we created an art form that combine so many different technical, mathematical, scientific methods of processing information to create something that suspend our desbelifed and allows us to engage in another world " -Falcon min 5:30 best quote ever
“Polygon is also a 2D shape” my mind is blown
Gameranx > Polygon
I never even thought about polygons much until I tried my hand at 3D modeling. Ever since I've had a lot more respect for people who do that all the time, it's really difficult and I didn't end up having a knack for it despite trying to get into it for a long time. It's kinda a thankless but vital job a lot of the time, cause anything you create that's cutting edge in the field today is gonna eventually be looked back on by people who saw what came after as if it's not a marvel you were able to do something like that when you did it.
There is another way to create 3D models and is much more realisic, having full volume, and they are voxels which is 3D pixels vs polygons
im so proud of humans for making this.
Linear Algebra
Fractal geometry has made such a profound impact on videogames and movies, great video.
You guys have some of the best videos on CZcams....but this one...this one might be your best yet.
I know it sounds weird, but i cant get enough of your voice Falcon...
Moon Knight
so its apparently gay if he likes someone's voice? and also, what's wrong with being gay?
Whoa this guy just opened my eyes.
this video is good for understanding polygons because its short,simple and informatinal to the point where it would be a good way to inform alot of people something a bit hard to understand in a simple way and it can also be used to understand polygons for math class for instance
There is an alternative to texture + polygon rasterization. It's called raycasting. It uses "voxels" (volumetric / 3D pixels) to represent objects.
Each voxel can be colored or transparent, and each voxel _can_ contain another 8 voxels. Those structures are called octtrees.
So unlike pixels in 2D pictures, voxels do not all have the same size, different levels of detail are possible.
An advantage of raycasting is that the necessary processing power for displaying a frame on the screen increases with the number of voxels displayed, but this number cannot get higher than the number of pixels on your screen. That means at some point, your computer wouldn't have to be faster to play more detailed games.
For this reason, you also don't need techniques like LOD (level of detail), since far away objects automatically consume less processing power since they cover less space on the screen.
Raycasting has other problems of course. One thing is that it's only computationally more efficient than polygon rasterization when there are really many details. (Unfortunately I don't know how many and if we are there yet.)
Another thing is that voxel models tend to consume much more disk space than polygon/texture models, since textures can be strongly (lossily) compressed, while lossy compression for voxel objects is more complicated.
I also heard that animations and dynamic lighting pose problems. Also, anti-aliasing needs very different algorithms. The whole field seems underdeveloped.
I'm a strong independant woman and I don't need no polygons.
JimMcnugget pixels for the win
Polygons on the screen are made of pixels.
AtotehZ
Pixels are sexist!
I'm a triggered polygon!
No matter what kind of woman you are, you're still a *minion* to the patriarchy *XDXDXDXDXDXD*
Wow i have been subed to this channel since 2012
Fusion Star nope
*Channel was created in June 2016* 🤔
This is true to an extent where all (at least since some time now) video game cards have dedicated math processors for the geometry related calculations. That's what gives them the "acceleration" designation for certain features.
Congratulations for the accessible explanation, thanks.
For those still wondering why triangles are used, it's just that they are the simplest abstraction: The simplest polygon, and conveniently, all other polygons can constructed from them. All the manipulation is done with simple linear algebra (translation, rotation, scaling) and trigonometry.
2:54 hey that’s blender
1:53 Video gam
6:15 Te- technically
Seems like you hit a home run on this one falcon. "Full time Game Developer" I like that you didn't leave out shaders, they are very important.
I do 3d art as a hobby and shaders are my favorite thing to work with! You have to think of it in terms of physics that relate to the real world (light distance, reflection, displacement, specularity, luminosity, transparency, etc.) then there are other purely digital aspects that trick the eye even further, like adding a sense of geometry to a polygon's face with bump or displacement maps, but that is only a fraction of the literal thousands of elements that can potentially go into a shader so that we can simulate the appearance of a surface found in reality within a 2d or 3d image. That being said my favorite types of 3d game and art is actually the hyper-conceptualized type. Like Fable, Warcraft, Final Fantasy. I prefer when the aesthetic doesn't strive for photorealism, but rather more creativity within the limits of that medium.
Blah blah GRAPHICS DON'T MATTER
When people say that I point them over to gameranx for class.
Legendsfaith3 Gameplay > Graphics.
Graphics don't matter in the sense that you don't need 100,000 polygons per sq. in. to have a good game. There are tons of 64-bit games that are still just as enjoyable as the games that are coming out now. I actually know people that won't play games that are on par with 7th gen or better; even then they are selective. So, yes, graphics matter, but not in every sense.
60fps>>>>>>>1080p+
It is still matter but if the gameplay sucks then it's just like purchasing a game without playing it or just watching scenery.
When it comes at the cost of optimization and literally have little to no (or Copy-paste) Gameplay, it's not worth it. But the gameplay is another discussion since nowadays you see Simple games being hardcore and complex feature-packed games being as basic as Atari but with tons of filler.
FALCON FOR PRESIDENT 2020 AHHHH
A few years ago we stitched together about 3 minutes of hd footage to create a 3d color model of a utility knife on a bucket. It took a lot of time to render, but it worked out pretty well. The end goal was to eventually create a "real" 3d model of a person to animate. Or "scan" real objects into a digital form to manipulate in maya. We had a 60 node backburner farm at the time to play with. Oh we also wanted to be able to 3d print crap we scanned in. We never got it down to a science because the students I was working with graduated.
While they were in their Ivory Tower talking about polygons I was in the streets making it happen every single day. I didn't have any fancy polygons, no, only my word and these two hands. I kept people entertained way before Super Mario and way after. Keep your polygons, I have my self-respect.
Nice channel. I enjoy it alot.
Another philosophical inshight: take it further... no finite process could simulate a real curve or store it through the steps. It always has to be stored as a list of points and a method of interpretation. I recommend Heidegger´s "Space and art".
What if we ourselves are living in a virtual space 😱
no
the matrix is real O.o
FoxerMapping could you imagine the amount of murders and suicide that would occur the hour after everyone found out that we live in a virtual world lol. So much rage quitting would immediately follow.
thesinaclwon Why? Just because it's a virtual world doesn't take away the fact that we only have one life, and that would confirm that nothing happens after death so the suicide rate would probably decrease
I don't think it would change anything
Video games are magical. Even having personally placed each point in my own games, I still get lost in the world and forget about the polygons.
There's an Australian based company called Euclideon that's been developing a new game engine that actually uses dots much like the real world is composed of atoms. The major advantage of this, apparently, is you can have extreme high quality graphics without compromising performance because you can use an unlimited amount of dots without a change in performance.
But can it run Crysis?
Ikr? Why not cubes?
FreePenguin5000 cubes are made of 6 polygons
WarioGiant 😐
FreePenguin5000 what?
WarioGiant I didnt know that so I made a 😐 face
FreePenguin5000 well think about it: polygons are just any 2D closed shape and cubes have six sides made of squares which are 2D and closed and therefore polygons
I've never thought about how a sloped curve could be continuously simulated into rendering. I'd love to see anything that can come from this.
Ever since I first played a 3D video game on N64, I've always imagined that someday computers will be powerful enough to render 3D objects on a microscopic level without the need for polygons. Maybe not molecular, but small enough that everything could be rendered with dots and each dot could have its own physics and collision detection so you could render things like individual grain of sand that a character could realistically run their hand through or walls that break precisely where you hit them. Basically something like voxels but on a scale so small that you can't see them. We're already getting close to that level of realism with a combination of polygons and other visual effects trickery, but we haven't quite reached the indistinguishable real-time reality that I imagined as a kid. I have no doubt that we'll get there, though.