Small retard question

Small retard question

Is clipping a problem with the "graphics" of a video game?

I understand that its from the physics engine that model collision comes from, but is it still considered a problem with graphics because its visually distracting?

Did you come from neogaf?
Personally idgaf about clipping, every 3d game has it

That's not what he asked you faggot

No I come from Sup Forums and /lit/

I never come to Sup Forums but my friend was confused about this point and I didnt have an answer for him

There are techniques that could prevent clipping such as using the meshes from the models for collision instead of simpler shapes like boxes or spheres. However, such a thing would be impractical since it would it is way more expensive in computing power while bringing very little benefit.
While you could say that this is a problem solved with computer graphics techniques, it's not really a problem and more of an acceptable innacuracy.

I never watched this anime. Is it true it turns people into pedos?

Speaking from experience, yes.

>it's not really a problem and more of an acceptable innacuracy.

My question is, is it correct to say that this is an acceptable inaccuracy for graphics? Is it wrong to say that this is "graphics" at all?

Sakura is very cute

You're born a pedo.

This is a case of the asker not being able to ask a question informed enough to give a proper answer to.

Clipping is when two 3D geometric shapes intersect each other. It happens in a multitude of cases. Physics and collision simulation is not a part of graphics engine. All graphics engines do is take a bunch of triangles, arrange them into a 3D shape, and project them to a 2D screen, no questions asked. It's up to the programmers, artists and the physics engine to make sure that there is no "clipping", wherever it is not desirable.

The most noticeable and pervasive one is when an animated mesh clips with itself. Like when a space mareen raises his arm and his giant pauldron clips through his head. That is because nobody does for a mesh colliding with itself, because that would be a pain in the ass and interfere with animation. Animation isn't part of the physics simulation, it's just a bunch of unconditional predetermined movements. It's the animator's job to make sure that his mesh doesn't collide with itself during animation. Of course, that isn't always possible because of modular design. A model of a human might not clip with itself when naked, but put a giant pauldron on it, and raising the arm will cause clipping.

Another type of clipping has to do with physics and collision simulation. There are cases when the physics engine itself fucks up and something phases through something that it shouldn't, but in most cases it has to do with manifold collision meshes used to calculate physics being different from the actual mesh itself. You see, if you tried to calculate physics for every single triangle of a hi res model with hundreds of thousands of polygons, computational power would run short. So what physics engines do, they have a much lower res model underneath for actually calculating collision. So you might use a hexagon to simulate a circle. Obviously, since a hexagon doesn't completely cover the area of a circle.
Dumb sakuraposter.

In the context of games. Yes this is an acceptable innacuracy since mantaining real time rendering is a must. In other contexts (such as rendering for movies or maybe in a software for designing buildings) this wouldn't be passable since speed isn't the focus, but accuracy of rendering.

>Is clipping a problem with the "graphics" of a video game?
I'd say so, it detracts from the overall visuals of the game far more often than not.

I want to commit a crime

Kind of, but not really. A game is ultimately rendered by commands telling the GPU "put this model here". Where the model goes is up to the developer, whether they handle it with animation or a physics engine.
Simulating physics can be harder than rendering the graphics in the first place, which is why developers simplify things. Thus you get clipping, but in a well designed game nobody will notice.
As an addendum, consider also that games are rendering simplified representations of objects, not totally realistic representations of them. A cloth on a table, or a patch of skin, will need special attention to even begin to attempt to emulate the physical properties of the real item in question.

Sakura was my childhood crush. There was never any hope for me.

Your confusion comes from the fact that you don't understand the difference between failure of tech and failure of artist.

Clipping isn't a problem in graphics. You tell the graphics engine to draw something, and it will draw that, unconditionally. You might ask it to draw two meshes intersecting each other, and it will do it, because that's what it's programmed to do.
If they DON'T want such behavior, it's up to you to prevent it.

Also, clipping isn't always "bad". Sometimes you just want meshes to intersect. Early 3D animations were done by having each limb as a separate model, and having them intersect at the joints to save on polygon count.

I love it when some redditor white knight starts claiming that pixels should be treated as real, living, breathing humans. How cucked can you be

>try to help a dumb weeb retard be less of a dumb weeb retard
>proceeds to shit up his own thread with inane autistic "smalltalk"
>instead of an engaginc discussion about 3D graphics technology, we get real time gelbooru simulation
Every time.
Never help a weeb retard, they never appreciate it.

Actually it's mostly newfags from Sup Forums, wich is reddit anyways.
Why did you even bring it up though, thread is comfy no one's complaining.

Back to Sup Forums

what I don't get is when there's blatant clipping or animations where two interacting models move completely out of sync

like if there's a single lever model (with different skins maybe, but the same shape in the end) that can be pulled, and the player model is the only guy who pulls them, how come the pulling animation so often looks completely out of sync, like the player character is acting in front of a green screen where he's told there will be a lever edited in. Why aren't the two models animated alongside each other, it seems like it would be easy as fuck and require minimal forethought.

It depends on the context.
If the clipping is happening during an animation (for example, a ponytail clipping through the body as the character leans forward) then it would be an animation issue.
If there is clipping when the model in question is in a neutral pose or their "standard" position/idle animation, then it would be a modeling issue.
If the clipping is occurring because of player interaction, such as seeing a second player's gun clipping into a wall because the player is running into it, then it's more of an unavoidable gaffe that isn't very feasible to fix without wasting time and system resources.
If the clipping is occurring because of a bug, such as a model not displaying properly, the object in question being rendered incorrectly, the object in question being off from its intended location, etc., then it would be a programming issue.

The latter two are usually considered the more "acceptable" forms of clipping, though if they occur too often the game likely had poor/rushed quality control and needed more polish.
The first and second examples are graphical and are less common in properly made games. They're generally considered more egregious, because avoiding clipping in models/animations is one of the most important aspects of 3D modeling and animation. As a recent example, Trails of Cold Steel features a Saber clone with very long hair. The hair is not animated, so anytime not-Saber turns her head slightly downwards, her hair clips through her breasts. This occurs very often in close-up cutscenes and is one of many lazy 3D animation gaffes that the game was mocked for upon release (in Japan; American players were bafflingly accepting of the laughable 3D animation in the game when it released overseas).

Video games are scripted to act a certain way and often times you're not meant to see those blemishes.

I don't understand what your point is. It's not some rare interaction here, I'm talking about two objects being terribly coordinated despite only really interacting with each other. Why can't they just slap a lever into the animation file for the player, then take that exact position data and put it in the levers animation file so it doesn't look like the player is holding air?

Everything looks a-ok here

Ok.

Say you have an animation of a human walking up to a lever, and doing a pulling motion with his hand. There's probably about a hundred thousand "frames" of animation (it's interpolated, but since it's happening real-time, we can call them frames).
Then you have the separate animation of a lever moving down.
Mind that those two objects aren't colliding on physical level or interacting in any way at all. All the game logic happens elsewhere in the code.
Now you have to sync those two animations. You have to time it when the human animation starts, and on which exact frame out of a hundred thousand, the lever animation should play.
All you have is the global time variable, animation start time and number of frames.

It's like taking two full length movies, hitting play on one, waiting a bit, and then hitting play on the other in such a way that at SOME point in the first movie, the dialogue syncs with the mouth movement of a character in the second movie.

Then you discover that the actor's arm moves at a different speed than the lever's animation, and the task was impossible in the first place.

Another question where the asker is so ignorant they don't even know what they're asking.

What do you mean by "file"? Files just hold data.
How is having two objects in the same file going to help you sync an animation in real time? What do files even have to do with this? Files are read into memory during the loading screen and then discarded.

What do you even mean by "putting the animation file into another"? How do you think it all works?

Also, there is an important distinction between model clipping and object clipping.

When we refer to models "clipping" through other models, it's a graphical error. However, it usually doesn't entail the actual object "clipping" through the other object within the game's engine itself.

To use the gun-through-a-wall example, the player is running into the wall and following the rules of the game, but their gun sticks out slightly from the player "object" collision box, which means that it appears to go "through" the wall. In reality, what the game is seeing is a different from the player. The player sees a man with a gun running at a wall with his gun in the wall (let's say the wall is thin, and the barrel is protruding from the other side of the wall). Logically, were he to fire his gun in this position, the bullet should exit the gun on the other side of the wall, but it doesn't. This is because the player character isn't actually clipping through the wall within the context of what the game sees. Since using model-perfect collision consumes significant system resources, it's not often used in games. When it is, it leads to disasters such as Dark Souls' Blighttown. Thus, in this example, the game simply sees a large rectangular box up against a wall. When the rectangular box (player) fires his gun, the bullets exit from his box and hit the wall, which he is directly against. From the wall hugging player's point of view, they're seeing effectively the same thing the game sees, as first person viewmodels are usually set to always appear on top of what the player sees, therefore they simply see their gun shooting against the wall, rather than sticking through it. This method itself can also lead to some unintentional gaffes whereupon getting too close to an object makes it look gigantic relative to your first-person arms and weapon.

Yes.

The simple fix to this is, while animating the player's lever-pulling animation,attach the model of the lever to the player and then animate them pulling the lever according to the lever's animation. This makes it easy and simple and should lead to little to no errors or roughness in the way the animations play. Really it's just laziness or engine limitations that prevent this method from being used.

I know they're separate, and that's the problem. When animating the things you should coordinate the lever and player with each other. When the animator is sitting with the pulling animation, he should add the skeleton for the lever into the animation and animate it alongside the player, so it doesn't look like they're holding air. It seems like it should be trivial to sync these if you're actually animating both at the same time and then splitting them up into two separate objects, if the animation itself matches, all you need to make it fit like a glove is to make them start at the same time, and snap the player model into exact position.

In reality, the guy and the lever were animated by two people who never spoke to each other.

Yes.
Accept it.
Become one of us.

this
or they thought about lever animation late down the cycle and the animator just did a quick job

Are you talking about inverse kinematics or something?

That doesn't eliminate the problem, just trades one set of problems for another.
Without mentioning the fact that it would be exponentially more difficult to coordinate objects in an already arranged 3D space instead of just having models move in predetermined locations, it would cause other visual bugs.
Such as characters teleporting slightly forward to "mate" with the lever, their hands could stretch out unnaturally, etc.
If something moves the actor out of the way while the animation is happening, something unpredictable could happen.
It's also a LOT more work.

Another problem this adds is splitting the workload for a simple animation between an animator, programmer and level designer. Instead of having an animator make two animations and save them to a file, you'll have to have special code for the lever, special code for the hand bone of the character models, the level designer has to use special lever meshes, and the lever itself has to have special code attached to it.
Instead of two objects with a bunch of animations, now you have a logistical nightmare and a polluted codebase where every object has scripts attached to it, with all the bugs that come with that.

It's exponentially more complex.

She was my gateway drug into loli, so there's probably some merit to that.

Do not sexualize the Sakura.

>Are you talking about inverse kinematics or something?
no, the computer animation equivalent of drawing the a guy pulling a lever on the same piece of paper and then cutting out the lever and putting the "being pulled" animation in its own file (to attach to the lever object), instead of drawing a guy pulling an invisible lever and also drawing a lever being pulled by an invisible guy on two pieces of paper

and the only problem would be snapping but that's less jarring than the alternative and is sort of required already if you want to make it look remotely like they're actually interacting with each other instead of the player just vaguely waving in the direction of the lever

In reality, all a game has to do is communicate its intent and gameplay mechanics to the player.
A character does the "crouch down" animation when picking an object off the ground ONLY to inform the player that a gameplay significant event has occured. That's all it needs to do.

It doesn't need to simulate the actual physics and body language of picking something up, it only needs to give gameplay tips to the player.

Guns in video games have muzzle flash and recoil not because it would be realistic, but because it gives the player feedback that something just happened.

Trying to simulate reality when the task at hand is to just make the game fun is a fool's errand.

you also can skip the lever+player animation completely and save the dev time. only showing some flash and lever changing the stance

Snapping into place to mate with the lever is far less jarring than FIRMLY GRASPING IT and watching the lever stab through the hand before sliding in front of the hand since the animation is a few frames ahead of the character's.

But that doesn't solve the problem.
A character could decide to pull a lever anytime, anywhere in the game and you still have to sync them.

Or maybe you have a different character with his own set of animations, with slightly different timings? Will you make another lever animation just for that character?
What if you have like five different races with their unique models in your game? Should there be a separate lever animation for each of them?

Should there be five different animations for every objects in the game, per each race / character model?
Just imagine, every time you add a dingy little doodad to the game, an animator has to do 5x times the work, making five different animations for each case.

I'm not talking animations where there needs to be a lot of variation, I'm talking one object only interacting with one animation. Or inanimate objects used in the animations of characters yet wobbling in the air in the vague area where the hand is held. Like if your mentor is dying in a cutscene and hands you his sword, the animation is obviously unique but it was done without considering the object despite that one not being animated at all.

It's probably more of an issue of trying to render a cinematic scene on the game engine instead of just doing it a pre rendered CG.

Coordinating 3D objects in real time interactive environments is a logistical nightmare desu.

>This is a case of the asker not being able to ask a question informed enough to give a proper answer to.

Im essentially asking, what is graphics for video games, and does clipping fall in that category as a problem for it?

You understood that on some level, because your answer is
> Physics and collision simulation is not a part of graphics engine.

Meaning your answer is no, so my question was clearly informed info for you to understand it, regardless if your answer is correct or not, it is relevant to what I asked meaning you understood what I was asking.

>Your confusion comes from the fact that you don't understand the difference between failure of tech and failure of artist.

>If they DON'T want such behavior, it's up to you to prevent it.

My question is, is clipping part of "graphics", specifically a problem. Your answer was no, because you can tell the graphics engine to not do that, but someone didnt have the foresight to prevent it from happening.

What is happening, is it still considered "graphics", thats what im asking. Say a game has a lot of clipping going on in the bad way, would you say this game has poor graphics?

>>proceeds to shit up his own thread with inane autistic "smalltalk"

I literally didnt do this. Two people asked me questions that didnt have to do with my OP and I answered them and it ended. That isnt small talk, and it didnt shit up the thread. The entire rest of the thread is relevant to my OP

What is wrong with you

Thanks, this helped clear that up. The particular instance we were talking about was a replay in nba2k18 where two players faces went into each other

>redditor white knight
ok, buzzfeed

You can say it's an animation problem.
Since both animation and physics simulation are acts of animating graphics.

"Graphics" is purely a static concept, it has to do with quality of still images.

Dumb weeb retard.

Pedos need to be gassed desu.

Are those shiny spots the tumblr nose of hentai?

Animation isnt part of graphics? Isnt animation just a subclass of graphics, which makes sense since you want to call it animating graphics?

>dumb weeb retard

I opened this thread saying the question was small and retarded so you arent saying anything but, but thanks for the answer anyway

>not liking a well-oiled loli
>not wanting maximum performance

A more important question is, are weebs a subclass of retards?
Sine I want to call you a dumb weeb retard?

Shadows.
Only, Z-fighting with the surface upon which they are cast.

Why is this so common?

This goes without saying, grown men larping as little girls can't be that smart.

Wouldnt the better question be, what do you call the retard who wastes their time asking himself, am I being redundant calling a dumb weeb retard a dumb weeb retard while talking to a dumb weeb retard?

You waste your time talking to a dumb weeb retard. No answers can help you, you fool

How cucked out of real life could you be touching your dick to pixels depicting loli. You'd have to be several notches below a fat, ugly, greasy pedophile.

>cucked out of real life

Youre trying too hard

There are no smart weebs so calling them dumb or autistic is redundant.

>one half of the thread is having in-depth discussions about clipping
>one half is sakuraposting
>and one half is flipping out and calling the sakuraposters pedo weeb dumbasses

Floating point bit depth is limited, duh.
You can't have an infinite number of fractions between 0 and 1, if the camera is 0, and the furthest point in the 3D space is 1, some objects between them are going to intersect, and the computer will just render the pixels of each one at some intervals.
It can be solved by just rendering the shadows into the textures, but it's more costly and especially more costly for real time dynamic shadows.
In that particular scene, however, I don't see why they didn't just bake the shadows in, it's just a side of a hill facing away from the sun.

My altruistic desire to help and inform even the dumbest of weeb retards outweights my burning disdain for them.
I'm just being "tsun" as you weeb retards like to say.

chill out

You wouldnt know tsun if it anta baka'd! you in the face

>faps to 2D
>not being cucked out of real life relationships, instead doubling down due to sour grapes and calling all women whores and 3DPD
>weebs telling others they're trying way too hard to fit in

Daily reminder that Sakura is a baka.

>he's still trying

I didn't even read what you said

Why would you be here shitposting on Sup Forums all day for over a decade, instead of being a productive adult, user?

The butthurt is real. Now enjoy your ban.

I already am a productive member of society but it's my day off and my gf is hogging the PC to play Sims so I have to phonepost.

You see, I have a need for social interaction and closeness, but I don't want to admit to myself that I am, on a level, dependent on other people for happiness, my ego is too massive to swallow that. Especially when it comes to dumb weeb retards.

So to even out the score, I insult people while also helping them, so they don't get the wrong idea that I enjoy talking to them or anything.
Being aware of this doesn't help me in any way to do anything about it. That's psychology for you.

>ban
You seem kinda lost.

2/3

A decent attempt

I believe you user.

What exactly are you trying to prove here, butthurt-sama?

Animation, physics and graphics are completely different things.
Clipping is a physics problem. The graphical part only is transcription of colors/textures and 2D projection, it has nothing to do with clipping

No, sakura is cute but she doesn't have "sex appeal", want a loli that actually turns people into lolicons? See Illya.

You cant have my boipuss Asuka

>tfw clipping will never, ever be gone

This, Sakura is a whore that laughs at his future boyfriend's penis being small.

>tfw there will always be network latency because of limit of light speed
You will never play competitive FPS with a person 12 timezones away from you.

Well this seems at odds with what others were telling me

I understand that clipping has to do with physics, but I dont understand how little or how completely different animation, physics, and graphics are.

Graphics is when you draw a circle.
Animation is when you move the circle around.
Physics is when you have two circles touching each other.

If you draw two circles clipping, you have fucked up the graphics.
If you animate two circles, and they clip, you have fucked up the animation.
If your two circles clip while touching each other, you have fucked up the physics.

This isn't that hard.

Clipping is not a graphics, physics, animation or whatever other category problem. It's a visual artifact.

The difference between an artifact and a problem is that an artifact is not necessarily problematic (it may occur and just look fine as far as the game's concerned).

Similarly it is visual and not specifically graphics-related because because it's an artifact you see but the cause may be any of a number of underlying things from any part of the program, and solutions to it may similarly come from any domain.

Clipping occurs because two triangles partly share the same space and z-buffering accurately represents that. depending on context and what caused it in the first place there are lots of solutions, e.g.:

-graphical solution: turn off z-buffering. No clipping ever anymore but space is no longer accurately represented (the PSX way). More subtle solutions include clearing the z-buffer instead of outright disabling it (the solution usually used for hands and weapons in FPS) or manipulating the depth calculations so that triangles that overlap in space do not in the z-buffer (seen in Guilty Gear Xrd, but only works in very specific kinds of layered scenes, ie you can't really use depth in a 3d sense anymore)

-gameplay solution: objects can never be close enough that they intersect. May be impractical, especially for articulated objects that have limbs moving all around because they'll need huge "bubbles" around them.

-physics solution: accurately represent the objects physically down to individual triangles. Not only is the intersection part computationally really expensive, it may actually be at odds with good gameplay, ie in an FPS you don't want your elbow to get stuck on every little thing and so on, and for both reasons collisions use approximations of actual objects, that may not even be the same depending on the kind of collision.

(continued)

If a movie has poor editing, does it mean it has poor cinematography?
The answer is, those are different things you retard.

Not all of those solutions are graphical, none of them will universally solve your problem. It'll be a combination of them, *including the notion that the clipping may actually not be a problem as far as you're concerned*, balanced to give the most appropriate result (which depends on your primary goal, could be hiding all clipping, hiding most clipping, being cost-effective), that'll solve your specific problem with it.

The answer can even vary, like you may want to prevent it for portrait shots of your fighting game character but can tolerate it during fights.

So, clipping is a visual artifact, with multi-disciplinarian and context-sensitive ways of preventing it. Of course that's not the convenient blanket answer you were looking for, but dev is like that.

Goddamnit she's so cute.

Person with actual knowledge here.

Most of the time, clipping has nothing whatsoever to do with physics. There are no physics involved when a character with a dress puts her arms down and they clip through the top of the dress, for example.

Clipping is in most cases the simple result of the fact that not every little piece of a 3D model is rigged to be animated. This is purely for cost/efficiency reasons. Why bother adding a special animation bone in every piece of garment that sticks out so it can be animated to fold, when it has so little effect on the quality of the game?

So, clipping is generally a "fault" with the rigging.

In cases where "physics" are involved, such as simulation-animated hair clipping through the shoulder, the clipping results from devs not wanting to sacrifice computing power to enable collisions between independently moving polygons. That can be very resource intensive, and if the polygons end up in a weird position they can easily end up glitching out and flipping all over the place, making it look even worse than with clipping.

> Of course that's not the convenient blanket answer you were looking for

I didnt say I wanted any particular kind of answer besides the right one

I mean I did get a "convenient" answer from this user if theyre right

I also bully my little brother like that.
I once put his legs through a plastic beer box mesh and he had to stand still for the whole day until parents arrived. That was hilarious.

Physics do cause clipping, retard.
Often in cases when the manifold collision model doesn't match the actual model itself.

Okay, what about the example me and my friend were discussing? Two models of basketball players in nba2k18 have their faces meld together

>Most of the time, clipping has nothing whatsoever to do with physics.
Clipping into walls because of simplified bounding boxes that aren't 1:1 to their graphical representation is extremely common though.

Collisions aren't physics, retards. Get educated before you try to talk.

Collisions CAN be physically simulated, but examples such as clipping slightly into walls when colliding with them have absolutely nothing to do with physics.

Collision IS physics, dumb retard.

What do you think physics objects do when they interact? They collide with each other.

Collision boxes for game objects are just hard coded convex hulls.

Whether you "simulate" it yourself or use a third party library, collision is still literally physics.

Jesus christ.

I don't play sports games, so I don't know when, if at all, physics are involved. What you're talking about might just be shoddy baked animations, or it may be physical collisions in which case the clipping happens because the models use a simple spheroid for collision detection of the head.

Third party library? Physics objects? You are fucking retarded and you don't know the first thing about what you are talking about.

Retards on Sup Forums need to stop trying to argue back like little children when someone who knows their shit enters the thread to put them in their place.

A car in a half decent racing game is a physics object and has physical collisions, yes. A character in an average action game is NOT a physics object and physics are NOT involved in the character's collisions. Repeat after me: Collision detection is NOT physics.

Collisions are programmed events that trigger when bounding boxes overlap, ZERO physical calculations are involved. This is completely distinct from physical simulations. Hell, even as a small child playing early 3D games like Tony Hawk's Pro Skater on the PS1 I could easily tell when the non-physical character was replaced with a ragdoll (aka physics object).

And yet here you are and still don't know the difference.

Could you tell just by looking if it was an animation or a physical collision?

Because the way this user put it
I would doubt they are ragdolls going through commands

Fucking retard, collision boxes are nothing but simple convex hulls for interacting with the level geometry, literally the same concept as when using convex hull for any other physics simulation, whether you want the player collision box to interact with world physics objects or not is entirely a gameplay choice.

>still going
Honestly, acting like you're hot shit and thinking you know what you're talking about when you're dead wrong is just so fucking embarrassing even I am feeling awkward having this discussion with you.

>if theyre right
They aren't.

How are they not?

>OP is a retard for asking this question
>no one can agree on what the answer is