[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vr / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / asp / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / qst / sci / soc / sp / tg / toy / trv / tv / vp / wsg / wsr / x] [Settings] [Search] [Home]
Board
Settings Home
/3/ - 3DCG



Thread archived.
You cannot reply anymore.




File: 3934yu85yu4.gif (458 KB, 393x373)
458 KB
458 KB GIF
>mfw gamers start talking about ray tracing
thanks nvidia
>>
What's hilarious is how few of them actually understand the significance of what has been achieved. Fuck em. I hope the gaymers ignore the 2080ti leaves more for the artists.
>>
>>638792
>mfw you messed the last 25 years with ray tracing and suddenly the plebs are flooding in
>>
>>638792
10(ten)GIGARAYS
I don't need it, but I want the rtx so badly.
>>
>>638796
Shiet m8, you must be old.
>>
>>638799
or I just started early
>>
the irony is that you can make loads more money rasterizing on mobile without any giga rays. This is all marketing, Fuck nvidia
>>
>>638799
At age 5? I doubt it.
>>
>>638803
Indeed.
>>
I mean I mostly use vray and iray to render stuff, so I should expect at least iray to be ultra optimized for RTX, vray who knows with chaos group

The really good thing for me is the whole nv-link thing though, finally an "affordable" 16gb card with two 1080s and of course they gimped the 1070 with no nvlink support

>>638803
what do you mean by this?
like ARM servers?
would that be affordable for an entry level archviz renderist
>>
>>638815
>what do you mean by this?
develop your game / app for cellphones using vulkan or opengl es. Your market is enormous, the job opportunities if you want to join a team are there
>>
>>638792
It's fucking incredible how many of them understand literally nothing about it. Just look at /v/, they're completely clueless.
>>
>>638826
>cards starting at $750
>/v/

jeez
>>
Well you see the issue is that we need to RISE UP
>>
>>638794
I'm new to 3d
>inb4 reee newfag
What's the significance of ray tracing, exactly?
>>
>>638933
Ray tracing offers more realistic or physically correct renders. However it is very taxing on your hardware and cannot be achieved for realtime. The significance of this development is that we are getting much closer to realtime rendering with ray tracing. Which is a TREMENDOUSLY 'UUGE deal for both the game and film industries.
>>
File: 9a6.jpg (22 KB, 398x500)
22 KB
22 KB JPG
>>638939
>>
>>638933
https://youtu.be/frLwRLS_ZR0
Here's a fantastic primer video done up by Disney. Path-tracing is similar to ray-tracing, just at a bit more of a complex level. The underlying concept is still the same, though.
What this means for graphics is that we no longer need to rely on many rasterisation techniques that current games still use. Things like ambient occlusion, shadow maps and reflection maps can become a thing of the past, since with ray-tracing we can accurately simulate how light would transport throughout a scene.
>>
>>638794
>>638792

cringe
>>
>>638946
But can we finally run Toy Story in real time?
I would love to see a movie done in a way that you could look around the scene while the movie is playing out. It's always been a weird dream of mine.
To think that I might actually live long enough to see that is pretty cool.
>>
>>638968
A couple more years until we are there
>>
>>638968
We could run something that looks exactly like Toy Story in real-time
The problem is everything behind Toy Story is not only old and outdated, but also unoptimized as fuck
>>
>>638991
BS
>>
I remember watching the livestream where they presented the technologies. They showed some real-time rendering of spheres with different materials and as usual the gaymers were all like "REEE WHAT THE FUCK IS THIS BORING ASS SHADER SHIT I WANT THE NEW GAMES AND A PRICETAG!!!" in the chat.
>>
>>639005
that part was pretty boring he goes for 5 minutes effect by effect like if the audience at siggraph wouldn't know anything
>and here we have refractions >and here caustics >and here global illumination
>>
isn't Nvidia's implementation of raytracing gimped? as in it's basically rasterization with a few shitty raytracing effects
>>
>>639005
>expecting intelligence from gaymers
that's a mistake you'll only ever make once
>>
>>639005
I waited two decades to finally see a cornell box with real time RT on consumer hardware.

Finally on the brink of ditching the fake rasterization and baking shit.
>>
>>638794
>the significance
Oh neat, some very basic blurry raytracing can now be achieved in realtime using 2 GPUs, $10k each
>>
>>639005
Nvidia's biggest mistake here is dangling pearls before swine.

>>639177
Yup you are in the camp of not getting it. Stay in school!
>>
Is this technology going to help with rendering? Aren't there floating point imprecisions, i.e., reproducibility difficulties? Or are they so small that they're unnoticeable in final renders?
>>
>>639185
where the fuck were you? Siggraph wasn't that long ago! https://www.youtube.com/watch?v=9A81NeQgJFE
>>
>>639186
Yeah, sorry, I should've been more accurate, I'm wondering about rendering for production, where you need pixel-parity between CPU and GPU. As far as I know, (at least previous generations of) GPUs have issues with floating point precision (or something along these lines), which makes them unsuitable for tasks that need reproducibility. Perhaps this new architecture doesn't have that problem, or it's small enough for rendering.

I recall that RenderMan XPU was expected to give CPU/GPU pixel parity, but now Pixar seems to be backing off a bit from that goal, expecting XPU to be more of a help during lookdev than a participant in final renders. But I could be wrong -- (I hope so!)
>>
File: confused animu girl.jpg (39 KB, 374x347)
39 KB
39 KB JPG
>>639186
wait what the fuck he said this is unoptimized spoderman?
from the movie?
and it renders in ONE SECOND?
>>
File: 1485234345617.jpg (34 KB, 480x522)
34 KB
34 KB JPG
>>639185
>unnoticeable in final renders
>raytraced GI
pick one
>>
>>638817
The realities of indie game development are rather sad.

https://www.youtube.com/watch?v=WycVOCbeKqQ
>>
>>639203
I think he meant if the floating point errors would be noticable.
>>
>>639224
well he said himself that after he removed all the crap with stock unity assets the numbers looked a bit better and let's not forget this accounts only PC\steam market. dude above said mobile and I'm really inclined to believe since one of my bros works as a coder, they literally make retarded clones of match 5 for ios\android and they sell somehow!
>>
>>639203
>>639226
My fault, I'm making it unnecessarily complex.

When I re-render something on CPU, if I haven't changed any parameters, the outcome is always the same at the pixel level, down to enough decimal places. I'm wondering if rendering with this new technology can do the same -- that's what I meant by reproducibility.

I have to admit I don't know much about GPU rendering, perhaps it's already not an issue. Bue seeing how Pixar is seemingly backing off a bit from their goal of pixel parity between CPU and GPU engines... I dunno.
>>
>>639267
>When I re-render something on CPU, if I haven't changed any parameters, the outcome is always the same at the pixel level, down to enough decimal places. I'm wondering if rendering with this new technology can do the same -- that's what I meant by reproducibility.
That kind of reproducibility, if it exists in GPU rendering, would only be available on Quadro hardware.
>>
>>638968
Toy story 1 used no raytracing. Renderman didn't have it at the time.
>>
>>639258
Yeah of course, but you said it your self. You friend is making clones. Those tend to sell for a month and then are forgotten. A lot of coders like that will just rehash the same game over and over in hopes that gaming the store will net them a profit.

That it self is a sad reality.
>>
>>639177

Yeah, the same type of noisy raytracing we've been doing on CPUs for at least the past 5 years
>>
>>639287
Yup, I've seen it first hand. Mobile games are like pissing in the wind. There's money to be sure but I've never caught any.
>>
>>639005
They had a much sicker demo on the show floor. I bet most of those fuckers would still not be impressed by this, when it's really jaw dropping: https://www.youtube.com/watch?v=O1RTWwyIbGY
>>
File: 1519748301340.jpg (34 KB, 798x486)
34 KB
34 KB JPG
>>639298
>$1200 for this
>>
>>639298
>>639299
Indirect reflection limit: 0.

Really impressive, nonetheless.
>>
>>639307
the thing is the current consoles cant implement it, the next consoles it'll be too expensive to put in so it wont have it, so you'll have rasterization for a good nother 10 years in ALL games
>>
>>639314
What you are saying is a chicken and the egg issue. Gamers cannot play it now cause no games support it and game developers can't support it because no one has it.

At some point it has to be introduced. Right now is the best time for nVidia to do it since AMD has nothing planned for at least another full year and even then the outlook for them is extremely poor given their profit margins can barely cover the costs of new lithographic masks.

I get the $150-$500 price increase is insane from a gamer perspective but I think most professionals are looking at that like a steal once production software supports it.
>>
>>639314
Even now rasterization's still going to stick around. This whole new RTX concept isn't completely shifting to ray-tracing, it's merely layering ray-traced shadows and reflections onto rasterization.
>>
>>639314
they already start it
https://www.youtube.com/watch?v=7_VhGvUzA2k
>>
>>638968
well yes. but it doesn't pay off.

in realtime everything is possible just like in raytrace the problem is certain things drop the FPS by alot.
for example world AO,hair,realtime lightning,cloth etc.
making a realtime movie would be a logistical nightmare since you would have to optimize every single part. also the quality of the lightning and shadow won't be as good.
>>
>>639344
meh, doesnt change much.
>>
>>638792
Test
>>
>>639344
>making a realtime movie would be a logistical nightmare
Exactly why it's a dream and not a reality.
>>639325
Of course it will, even in 20 years, they'll still have lots of raster graphics. People don't like change. It's going to take awhile before we're completely on raytracing and even then some people will just continue to program the old way because they're relics.
>>
>>638792
Great way to sell an incremental update as something new. Now it's not about the CUDA shit but FAGGO-RAYS.

They were talking about real time ray tracing since 2004. That Asian guy is just reheating old sausages before they expire.

The only thing notable is AI-based noise reduction. But you don't need a dedicated chip for that.
>>
>>639423
>some people will just continue to program the old way because they're relics.

There is still a legitimate reason to use rasterization, and that is for stylization. If you don't care about reality, raytracing is a waste.

>>639427
>FAGGO-RAYS
Did you intend to be this cringy when you left your board for this post?
>>
>>639427
It's not really incremental when performance is 6x more in real-time ray-tracing than the generation that came out a year before it, is it?

>>639428
>stylization
Uh, no? Unless you're referring to NPR, there's really not much "style" that exists in the bleeding edge rasterization techniques we have today. Ray-tracing's simply just the next step to take after rasterization.
>>
>>639442
>Unless you're referring to NPR
I literally said if you don't care about reality in this post >>639428

More over there are still situations that rasterization just make more sense for. HUDs, overlays, Filters. Really do you think Text is going to be raytraced? Rasterization is going to be used by nearly every game, beyond 20 years, because it's efficient and makes sense under many circumstances. Raytracing makes sense when you want to simulate reality.
>>
>>639428
Even stylized stuff will end up being ray traced, but you make a good point.
>>639443
They could always go with SVG for things like that.
>>639427
>They were talking about real time ray tracing since 2004
Try since like 1986. Just because people have been "talking about it" doesn't mean it was released. Let alone to the public consumer.
Here's your (You).
>>
>>639466
>SVG
Scalable vector graphics? In a game? I don't see how that relevant.
>>
>>639472
>>639443
>More over there are still situations that rasterization just make more sense for. HUDs, overlays, Filters. Really do you think Text is going to be raytraced?
For this part of the post. SVG makes a UI future proof for higher resolutions. Some games have already done this by using flash graphics for their interfaces or menus.
>>
>>639005
"REEE WHAT THE FUCK IS THIS BORING ASS SHADER SHIT I WANT THE NEW GAMES AND A PRICETAG!!!"
Reminds me of the story where some developers were able to render a black triangle in realtime for the first time in history and the non-programmers were like "why are you so happy about this?" not understand the implications of how important that breakthrough was.
>>
>>639474
>For this part of the post. SVG makes a UI future proof for higher resolutions. Some games have already done this by using flash graphics for their interfaces or menus.
???
SVG is rendered using Rasterization. Depending on the game and planing involved some games will use Libraries like https://www.freetype.org/, Render all the characters they use to a texture at what ever resolution they need and then raster from that. They can also get a set tris the same way which again is render with rasterization.

SVG and raytracing are extreme opposites here. The format is not optimal. (Possibly with IRT maybe, but who uses those?)

Why is resolution relevant to this?
>>
>>639443
>>639478
You're really scraping the barrel in this non-argument here. We're all talking rasterization in the sense of 3DCG, and here you are attempting to extend it to 2D graphics because everything has to be an argument to you, I guess.
>>
>>639481
WTF are you talking about? I never once brought up 2D graphics.

Raytracing is for simulating real lighting, rasterization is for drawing what ever you want where you want it fast.

Look, these are some examples of what I'm talking about from >>639428 >>639443


Depth of field + plus 3d texts that's occluded by objects but doesn't interact with them. A bitch to implement in ray tracing but trivial with a depth buffer and rasterization.
https://www.youtube.com/watch?v=90gv5KtNZ4w

Detective mode has no barring in reality, you would need to re-write the raytracer to pretend its doing raster.
https://www.youtube.com/watch?v=Klo6xyC01iA

Okami...the whole game.
https://www.youtube.com/watch?v=RhnhRfTzRnw
>>
>>639273
>>638991

Theoretically speaking , you stupid pedants.

He means in theory could a production *like* Toy Story, (or ANY 3D animated movie) be created for viewing in real-time.

THIS is what I estimate will push VR to it's ultimate conclusion. Imagine going to the cinema or some specialized setting, wearing headset and being entirely SURROUNDED by the movie. That is the future. """Gaming""" is a meme that normies aren't interested in. Immersion is key.
>>
>>639532
>Imagine going to the cinema or some specialized setting, wearing headset and being entirely SURROUNDED by the movie.
yeah I think this is where it all goes, from first muddy pictures on the walls we always wanted to capture reality and dreams
>>
>>639538
This is why facebook, a social media company, bought Oculus. Because they are in the business of predicting the future, hell, even NOW people are addicted to their smartphones. When the devices are portable and affordable enough, your average normie will put on their headset, sitting with their family together to watch a TV / movie together, or play Farmville VR.

I see this as 6 - 12 years a way. The headset will be wireless, full eye/cornea tracking, smart voice recognition and preliminary advances in thought interfacing.
>>
>>639532
>That is the future. """Gaming""" is a meme that normies aren't interested in. Immersion is key.
What did he mean by this?
>>
>>639442
>It's not really incremental when performance is 6x more in real-time ray-tracing than the generation that came out a year before it, is it?

They could achieve the same thing by writing a better driver for the previous generation of cards.

Nvidia always has fake product tiers and fake press material. After buying the 970 with FAKE amount of RAM I will never buy anything from these cunts unless it's objectively reviewed by 3rd parties. They also don't make enough cards so they can sell units for more.
>>
>>639567
>They could achieve the same thing by writing a better driver for the previous generation of cards.
nice job pulling that one out of your ass, nobody knows what the raytracing cores actually do.
For all we know they could be running gimped FP8/FP16 ops at 4x/2x the speed of single precision float.
>>
>>639567
>They could achieve the same thing by writing a better driver for the previous generation of cards.
You clearly have no idea whats going on under the hood. This is DX12 and eventually Vulkan, There is very little driver involved.

>>639569
>For all we know they could be running gimped FP8/FP16 ops at 4x/2x the speed of single precision float.
Plausible, but I'm pretty sure that said they are doing some BVH optimizations, so I'd imagine they are doing work to deal with tree navigation
and ray management. But you're right, we have to wait and see.
>>
>>639478
>>639524
Clearly you meant something different when you said "HUDs, overlays, Filters." I thought you meant UI. (Which is why I said UI) SVG need not apply for what you meant.
Developments in raytracing (as well as other future technologies) might make the methods of old not worth using in comparison.
https://youtu.be/G01FZUtIMX0
Just look at the 3D games from 30 years ago.
Who knows what leaps we'll make in the next 30.
Even raytracing might be rendered obsolete by a more accurate/faster method.
>>
>>639548
Which is funny because all fuckerberg did by buying Oculus was drive it into the ground.
>>
>>639657
>Clearly you meant something different when you said "HUDs, overlays, Filters."
No, I was actually specifically thinking of Batman and L.A Noire. I clearly said stylization >>639428, You even quoted me on that >>639442


>>639657
>Even raytracing might be rendered obsolete by a more accurate/faster method.
This is physics simulation. That's only happening if we get new physics.
>>
File: You_win_the_prize.jpg (45 KB, 523x349)
45 KB
45 KB JPG
>>639657
>Even raytracing might be rendered obsolete by a more accurate/faster method

/3/ the post *golfclap*
>>
>>639694
>new physics
just in: madlad on /3/ demands universal laws to be rewritten
>>
>>639660
How?
They're still selling headsets that people are buying.
>>
>>639739
Not him (>>639660) But hype has really died down sense it was introduced. Hardware complexity, small audience and lack of software to support it really limit growth.

Hate to say it, but an all in one VR solution is pretty much the only way it stands a chance of mainstream adoption. Some thing like modern Tegra or Snapdragon 845 in an iteration of Windows Mix Reality would be best I think.
>>
File: 1451659381631.jpg (66 KB, 419x249)
66 KB
66 KB JPG
>>639744
>an all in one VR solution is pretty much the only way it stands a chance of mainstream adoption.
They already have the Oculus Go, which I don't think they really marketed.
>>
>>639532
>"""Gaming""" is a meme that normies aren't interested in. Immersion is key.
activates my almonds
>>
>>639694
>Mixing up (You)s
There are more than 2 posters on /3/ you know.
>That's only happening if we get new physics
But that's wrong you fucking faggot. Raytracing isn't a 1:1 simulation of the world. It can still be improved upon with new techniques that we have yet to discover.
>>
>>639748
The oculus go is an android phone similar to GearVr
>>
>>639739
It's more of >>639744 He didn't do anything to keep the hype train going. Or at least it didn't seem like it.
All he really seems to be doing is telling his investors to "keep waiting" until it actually starts to turn a profit.
>>
>>639868
>There are more than 2 posters on /3/ you know.

Let me get this straight.
>>Some guy started a conversation on how raytracing was going to replace raster. >>639423
>>>I said stylization was still a thing >>639428
>>He said not unless it's NPR >>639442
>>>I said yes exactly >>639443
>> Then he says raytracing will do stylization and brings up nonsense on SVG >>639466
>>> I say SVG? >>639472

>>Then for some reason, you, a different person, randomly interjects an explanation of what HE meant. >>639474
>>> I explanation SVG >>639478
>>You take offense to explanation and assume I meant 2D >>639481
>>>I reiterate what I mean with examples >>639524
>>You say some nonsense about better than raytracing >>639657
>>>I say that's not happening. >>639694
>> (you) >>639868

All while maintaining the same posting style.

Look, Raytracing isn't a specific thing. It is a way of going about simulation, relying on finding intersections in a scene and batching out more rays dependent on the intersections and depth. It is fundamentally particle physics simulation. There is no one set raytracer, they can all operate differently.

So it's not about finding something better to replace raytracing, but that raytracing is the technique used to simulate particle physics. Like i said, Rasterization will always be a thing, and Raytracing is something that can be used to simulate real world light.

>>639868
>>it can improved upon with new techniques that we have yet to discover.
I want you to think about this statement and why it's flawed.
>>
You incorrectly credited "his" and "my" posts several times. And "he" was also probably multiple people. Please learn how 4chan works before you post on it.
As for the rest of your post, please see the above if you expect to have debates on 4chan.
>>
>>639875
>Please learn how 4chan works before you post on it.
I'm guessing at this point you've figure out that I'm using a plugin that notifies me when I'm quoted. That or this is a very ironic post.


Care to enlighten which posts are yours?
At what point did me explaining the uses of raytracing, rasterization and SVG become a debate for you?
>>
>>639876
>I'm guessing at this point you've figure out that I'm using a plugin that notifies me when I'm quoted. That or this is a very ironic post.
Forgot to quote your stupid ass, sue me.
>Care to enlighten which posts are yours?
No, I don't. Learn how 4chan works.
>At what point did me explaining the uses of raytracing, rasterization and SVG become a debate for you?
In literally the first post you quote as yours in >>639873 where you debate on what the uses of raytracing are.
https://www.dictionary.com/browse/debate

Now if we can continue.
I'll let you have your "raytracing can't be obsolete ever." Why do you think rasterization won't be replaced? What styles don't work without rasterization?
I don't know much about NPR, but could it not be layered over top of a raytrace render method rather than rasterization?
>>
>>639873
>>639442 and >>639481 (and >>639325) are my posts. At least 4 different people in the whole conversation you quoted, you and me included. Writing styles are clearly distinct so you're pretty poor at identifying "same posting style".
I stopped my side of the conversation then because I realised you were talking pretty much out of your ass and some anon had already kindly tag-teamed in.
This post just here to BTFO you a lil' more :^)
>>
File: you-make-my-head-hurt.jpg (58 KB, 300x269)
58 KB
58 KB JPG
>>639880
>>639881
>Posted 6 minutes apart.
Fine.

Rasterization of triangles is done by mapping the triangles space to screen space via bresenham's line algorithm. The number one benefit is speed. At any give time you work with 1 set materials and 1 shader program. You also have the benefits of locality for cache and memory. You are writing to roughly same spot on the screen. This makes rasterization very fast.

Raytracing does not map. Every screen pixel must cast a ray that will iterate through a tree of the entire scene, accessing random areas of memory with no locality. So every object, every volume, and everything that you want the raytracer to draw must be in the scene. If you have objects in the you scene that are only used on the 1st intersection, every n+1 intersections will be slowed down by their presence. Ideally, you only want things in your scene that are part of the simulation.

Raytracing is inherently slower than rasterization and only becomes faster when the number triangles on screen explodes. However, rasterization has mitigated that issue with LOD.

Moreover, rasterization benefits by interpolating very evenly between the vertices of a triangle, enabling predictable noise, such as banding or moire effect. Those are dealt with by dithering and mipmapping.

Raytracing gives you no guarantees on where on an object you will intersect. The result is noise and lots of it. Two ways to deal with this are supersampling(casting multiple rays per pixel) or denoising algorithms, which are basically just more costly blurring functions. Both are extremely costly in performance.

So based on all of that, if you are not simulating light, it's faster to use rasterization. Rasterization is fast and easy enough to do that its worth doing than forcing a raytracer to do something it wasn't meant to.
>>
File: Screenshot_110.png (19 KB, 697x274)
19 KB
19 KB PNG
>>639887
>18 minutes is somehow 6 minutes apart
Posts have timestamps, y'know. Like I said, I'm already out of the conversation so you two can have your intimate moment together, but I still don't understand why you can't just accept that there can be several participants on one side of an argument.
>y-you could have just edited that image!
Yeah, like I'll waste effort on editing a screenshot.
>>
File: 121.png (216 KB, 568x1023)
216 KB
216 KB PNG
>>639889
>18 minutes
Don't know why I saw six. 18 is not much better.

Fine, you claim is yours >>639481
I reply with >>639524
To which >>639657 replied saying
>Clearly you meant something different when you said "HUDs, overlays, Filters." I thought you meant UI. (Which is why I said UI) SVG need not apply for what you meant.
Which claims ownership of >>639474 since he brought up UI.


But according to you, neither >>639657 nor >>639474 are you, despite responding to conversation thus far not apart of.
>>
File: 1458049438988.png (28 KB, 1004x520)
28 KB
28 KB PNG
>>639869
>an android phone similar to GearVr
It's not a phone, it's an All-In-One headset. Unless you meant to say that it's like a phone in the inside.
>>639870
>"keep waiting" until it actually starts to turn a profit.
That's a massive red flag.
Any chance of the hype train starting back up? It seems like VR companies aren't marketing at all and are making bad decisions in general.
>>
>>639887
>>639891
>Images
The feeling is mutual, trust me.
>Rasterization is fast
So, remove that speed. Aside from speed, what does rasterization do that makes it so special?
In the future when raytracing is so easy to do that the difference isn't even noticeable, what exactly is it that rasterization will be doing better?
>>639891
>neither these posts are you, despite responding to conversation thus far not apart of
It's almost like... there are multiple people in a single conversation. Fascinating.
Why are you so hung up on this?
>>
>>639895
>Any chance of the hype train starting back up?
The hype train only really took off for early adopters. Which is what they need to be doing anyway. Normies only adopt new technology once enough early adopters have given it the green light and made it obvious that it's good technology.
They're sheep that follow the 10% of early adopters/inventors around, but they're 90% of the market.
>>
>>640051
>So, remove that speed.
>Method A is faster than Method B
>Make Method A slower, then use Method B
Why do you want raytracing to be used for everything?


>>640051
>what exactly is it that rasterization will be doing better?
How would you use raytracing to perform motion blur?


>>640051
>Why are you so hung up on this?
The second I corrected you about SVG, you became extremely defensive. Like pointing out your lack of understanding caused you personal injury.

Basically, it's that age old advice. If you don't want to be bullied, don't show that it bothers you.
>>
>>640059
>Make Method A slower, then use Method B
Thanks for putting words in my mouth. The idea is that Method B will be "just as fast" as Method A eventually, so we have to look at the advantages of Method A.
If Method A's only advantage is the speed, then we have to make sure that there isn't something that we're over looking that would make Method B faster than Method A.
>How would you use raytracing to peform motion blur?
See now, that's a good point.
>The second I corrected you
Coming from a guy who can't even tell "(You)s" apart from one another.
I didn't mention SVG in that post. Are you sure you're not the one bothered by this?
>>
>>640060
And to clarify on poor wording before you nitpick that... No. Raytracing will never be faster than rasterization. But who cares if they both render before the computer is even calling for the next frame.
>>
>>640060
>Method B will be "just as fast"
>>640061
>No. Raytracing will never be faster than rasterization
You are underestimated how the methods scale.

Blender, well known app here in /3/, Is moving to a rasterizer because of how slow ray tracing is.


>>640060
>See now, that's a good point.
I explained many other including the noise issue.
I also noted the issues with scene. The entire scene must be loaded into memory.
When does Tessellation occur? Or do you attempt to use NURBS? Can they be dynamic? How do you store them in optimal manner? Tessellate them?
How deep do you traverse the BVH tree before you are certain there is no light source?
What do you do when composition demands impossible lighting?
How do you draw a 1 pixel wide line with a raytracer?

You still haven't answered: Why do you want raytracing to be used for everything?


>>640061
Could you stay out of this I'm trying to have a conversation with this guy >>640060 and you keep interrupting.
>>
>>640062
>Could (You) stay out of this, I'm trying to talk to (You)
(You) have lost it, anon.
No, I don't want raytracing to be used for everything. I don't even use raytracing. At this point I've mostly just been amused with how retarded you are and wanted to keep it going.
I don't know that much about raytracing, I'll be honest, but debate (yes, this is a debate) opens your mind to new ideas and you learn things you didn't know before.
The only thing I'm sad to say is that I might have actually learned something from your dumb ass.
Feel free to mark this as a win for your "Someone is wrong on the internet" card.
>>
>>640062
>Blender, well known app here in /3/, Is moving to a rasterizer because of how slow ray tracing is.
That's not what Blender devs are doing, and, anyway, nobody cares about what they do.
>>
>>639299
HURR HURR LEMME PIC THE WORST-LOOKING SPLIT-SECOND SCREENSHOT I COULD FIND JUST TO MAKE THIS THING I DON'T LIKE LOOK BAD ON FORE-CHANG





Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.