ADM-201 dump PMP dumps pdf SSCP exam materials CBAP exam sample questions

针对光模拟的实时噪声滤波 – 译学馆
未登录,请登录后再发表信息
最新评论 (0)
播放视频

针对光模拟的实时噪声滤波

Real-Time Noise Filtering For Light Simulations | Two Minute Papers #181

亲爱的学霸们 这里是Károly Zsolnai-Fehér带来的两分钟论文
Dear Fellow Scholars, this is Two Minute Papers with Károly Zsolnai-Fehér.
在这个系列里 我们谈了很多逼真图形渲染 它是
In this series, we talk a lot about photorealistic rendering, which is one of the most exciting
计算机图形研究里最令人激动的领域之一
areas in computer graphics research.
逼真图形渲染意味着我们把虚拟物体放到场景里 给它分配材料模型
Photorealistic rendering means that we put virtual objects in a scene, assign material
然后运行光仿真程序来创建一个美丽的图像
models to them, and then run a light simulation program to create a beautiful image.
这个图像描绘这些物体在现实中的样子
This image depicts how these objects would look like in reality.
这对电影工业非常有用 因为我们能创建非常逼真的场景
This is particularly useful for the film industry because we can create highly realistic scenes
以现实生活中不能做到的方式设置它们
and set them up in a way that we couldn’t do in real life.
我们可以拥有我们能够想象的任何物体 拥有我们从来就买不到
We can have any possible object we can imagine, light sources that we wouldn’t ever be able
的光源 改变一天中的时间 甚至改变我们居住的行星
to buy, change the time of the day, or even the planet we’re on.
我们几乎拥有无限的预算
Practically, we have an infinite budget.
那太棒了!
That’s amazing!
然而 创建这样一幅图像需要很长时间 经常需要大约几小时到几天
However, creating such an image takes a long time, often in the order of hours to days.
这里你可以看到我渲染一幅图像 甚至如果镜头被大幅度加速
Here you can see me render an image and even if the footage is sped up significantly, you
你可以看到它将花费很长很长的一段时间
can see that this is going to take a long, long time.
Spp表示每像素的样本数 也就是我们计算的每像素的光线数量
Spp means samples per pixel, so this the number of light rays we compute for every pixel.
spp越多 我们获得的图像就越干净 越收敛
The more spp, the cleaner, more converged image we get.
这一技术执行时空上的过滤
This technique performs spatiotemporal filtering.
这意味着我们输入一个有噪声的视频 并尝试消除噪声
This means that we take a noisy input video and try to eliminate the noise and try to
猜测最终图像的样子
guess how the final image would look like.
它能从充满极端噪声的输入中创建几乎完全收敛的图像
And it can create almost fully converged images from extremely noisy inputs.
如你所见 这些视频是以1样本/分钟创建的 它的噪声就和它所获得的一样大
Well, as you can see, these videos are created with one sample per pixel, which is as noisy as it gets.
这些1样本/分钟的图像能够极快地创建出来
These images with the one sample per pixel can be created extremely quickly, in less
每个图像只用不到10毫秒时间 这个新的降噪器同样花费大约10毫秒时间
than ten milliseconds per image, and this new denoiser also takes around 10 milliseconds
来从嘈杂的输入中重建最终图像
to reconstruct the final image from the noisy input.
是的 你没有听错 这就是最终的实时结果
And yes, you heard it right, this is finally a real-time result.
它能够实现全有赖于对直接和间接的光源效果的去耦
This all happens through decoupling the direct and indirect effect of light sources and denoising
和分别去噪
them separately.
与此同时 算法同样尝试预测图像不同部分的噪声数量
In the meantime, the algorithm also tries to estimate the amount of noise in different
以便为降噪例程提供更有用的信息
parts of the image to provide more useful information for the denoising routines.
事实上 运行在显卡上的整个流程是对算法简明性的
The fact that the entire pipeline runs on the graphics card is a great testament to
一个很好的测试
the simplicity of this algorithm.
无论你何时看到SVGF的术语 你都能看到新技术的结果
Whenever you see the term SVGF, you’ll see the results of the new technique.
我们有这些1spp的带有噪声的输入视频
So we have these noisy input videos with 1 spp.
还有……看这个!
And…look at that!
哇!
Wow!
这是那些看起来如同魔法的论文之一
This is one of those papers that looks like magic.
它没有用到神经网络或学习算法
And, no neural networks or learning algorithms have been used in this work.
不久之前我推测 或准确地说 希望实时图像渲染
Not so long ago, I speculated, or more accurately, hoped that real-time photorealistic rendering
在我有生之年能够成为现实
would be a possibility during my lifetime.
就在几年后 这篇论文出现了
And just a few years later, this paper appears.
我们知道计算机图形研究的进展速度是惊人的
We know that the rate of progress in computer graphics research is just staggering, but
但这个 这个实在是太难掌握了
this is, this is too much to handle.
超级兴奋看到艺术家们将会在哪里使用它 当然
Super excited to see where the artists will take this, and of course, I’ll be here to
我将会在这里向你展示最酷的后续进展
show you the coolest followup works.
感谢观看和慷慨支持 我们下次再见!
Thanks for watching and for your generous support, and I’ll see you next time!

发表评论

译制信息
视频概述

两分钟论文,针对光模拟的实时噪声滤波,实现逼真图形渲染。

听录译者

收集自网络

翻译译者

B11101001

审核员

审核团O

视频来源

https://www.youtube.com/watch?v=HSmm_vEVs10

相关推荐