您需要 登录 才可以下载或查看，没有账号？注册
x
本帖最后由 秋令6689 于 2022-5-17 12:11 编辑
MJ’s Room – Environment Breakdown – Roy Bennett
MJ的房间 – 环境分解 – 罗伊·贝内特
Intro
Hello! I’m Roy Bennett, an online student at Think Tank Training Centre about to enter my final term and mentorship studying Environment Art for Games. Getting into 3D Art was kind of an act of trial and error. I always knew I wanted to pursue a career in video games, but it’s such a vast industry with so many disciplines, it was difficult to see where I could fit. My current career has been in the IT field, troubleshooting computer issues and doing system administrator work—but after 10-years doing that, the desire to pursue this creative industry only grew.
So in an effort to pursue this passion, I figured the best way for me to start the journey was by going back to school. I did some research and found Full Sail University. They had a Game Art degree program, but at that time I thought I couldn’t do that—I mean, me, an artist? Pfft. I’m a tech guy, I couldn’t even draw! So instead, I went for the next best thing which was their Game Design degree program. It had all my boxes checked; it was creative, technical and allowed me to work with a team of classmates to create a video game by the end of the term.
However, during my time there I was introduced to 3D tools and game engines like Autodesk Maya and Unity. I’ll never forget rendering my first image in Maya—I thought it was the coolest thing on the planet. It isn’t anything special, but I saved that first render, it was of a tree, picnic table and a fence. That feeling of amazement grew to a curiosity of, “What else can I create in 3D?” That’s where I believe my interest in 3D Art began.
I still kept that voice inside silent though, believing I couldn’t do 3D Art and that I wasn’t an artist. So upon graduating from Game Design, my classmates and I released our first game, Duncan & The Wisp, on Steam and parted ways. All looking to our futures and what was next.
Some of them got jobs in QA and made their way into design positions, but I was still stuck wondering if design was where I belonged. A few weeks after graduation, I had thought about it and figured that my favorite part about game design was getting to block out levels which is the gameplay space players would experience. I thought maybe I could specialize in that and enrolled in CGMA’s Level Design for Games course. It was an awesome experience and I learned a great deal from Naughty Dog’s own Emilia Schatz.
Having taken a lot away from that course, I realized that I enjoyed detailing my blocked out environments with subtle details a lot more than the design process itself. I wanted to flesh them out more, add textures, detailed props, etc. Speaking with a friend about this he said, “Dude, you’re not a designer. You’re an artist. BE an artist!” which of course contradicted everything I have been saying to myself from the start.
Just be an artist. Okay. Why not? I felt I’ve come so far and tried a bit of everything already—might as well do the one thing I haven’t done, or maybe was too afraid to do.
Enter Think Tank Training Centre—I enrolled in their Online Environment for Games course and started doing the thing I thought I couldn’t and worked on 3D Art. Needless to say, it’s where I found my flow. Everything just started to click and I stopped questioning it—instead I just ran with it. I could fine-tune those details, push the quality higher, breathe life into the image and pour my passion into every project. After my first term at Think Tank, I was awarded a scholarship for my 3D rendering of Milqui, a concept by Alejandro Burdisio. I finally felt like I was on the right path and everything I learned throughout my journey so far was just preparing me to dive fully into 3D art without any reservations.
介绍
你好！我是罗伊·贝内特（Roy Bennett），智库培训中心（Think Tank Training Centre）的在线学生，即将进入我的最后一个学期，并担任研究游戏环境艺术的导师。进入3D艺术是一种反复试验的行为。我一直都知道我想从事电子游戏事业，但这是一个如此庞大的行业，有这么多学科，很难看出我能适应什么。我目前的职业是IT领域，解决计算机问题并从事系统管理员工作 - 但是在从事这项工作10年后，追求这个创意产业的愿望只增不减。
因此，为了追求这种激情，我认为开始这段旅程的最佳方式是回到学校。我做了一些研究，找到了Full Sail University。他们有一个游戏艺术学位课程，但当时我认为我不能这样做 - 我的意思是，我，一个艺术家？Pfft.我是一个技术人员，我甚至不会画画！因此，我选择了下一个最好的东西，那就是他们的游戏设计学位课程。它检查了我所有的盒子;它具有创造性，技术性，并允许我与一群同学一起在学期结束时创建视频游戏。
然而，在我任职期间，我接触到了3D工具和游戏引擎，如Autodesk Maya和Unity。我永远不会忘记在 Maya 中渲染我的第一张图像——我认为这是这个星球上最酷的东西。它没什么特别的，但我保存了第一个渲染图，它是一棵树，野餐桌和篱笆。这种惊讶的感觉逐渐变成了一种好奇心，"我还能用3D创造什么？这就是我相信我对3D艺术的兴趣开始的地方。
Bennett_Scene_FirstRender
不过，我仍然保持沉默，相信我不能做3D艺术，而且我不是艺术家。因此，从游戏设计专业毕业后，我和我的同学们在Steam上发布了我们的第一款游戏《Duncan & The Wisp》，并分道扬镳。所有人都在展望我们的未来和下一步是什么。
他们中的一些人在QA部门找到了工作，并进入了设计职位，但我仍然想知道设计是否是我的归属。毕业几周后，我想到了这个问题，并认为我最喜欢的游戏设计部分是屏蔽关卡，即玩家将体验到的游戏空间。我想也许我可以专攻这个，并报名参加了CGMA的游戏关卡设计课程。这是一次很棒的经历，我从Naughty Dog自己的Emilia Schatz那里学到了很多东西。
从那门课程中学到了很多东西，我意识到我喜欢用微妙的细节来详细描述我的封闭环境，而不是设计过程本身。我想充实它们，添加纹理，详细的道具等。在和一位朋友谈论这件事时，他说："伙计，你不是设计师。你是一个艺术家。做一个艺术家！"这当然与我从一开始就对自己说的一切都相矛盾。
只要成为一名艺术家。 好。为什么不呢？我觉得我已经走了这么远，已经尝试了一些东西——不妨做一件我没有做过的事情，或者也许太害怕了。
罗伊-贝内特-罗伊-贝内特-最终-小规模
进入智库培训中心 - 我参加了他们的在线游戏环境课程，并开始做我认为我不能做的事情，并从事3D艺术工作。毋庸置疑，这是我找到心流的地方。一切都开始咔哒咔哒，我不再质疑它——相反，我只是带着它跑了。我可以微调这些细节，将质量推得更高，为图像注入生命，并将我的激情倾注到每个项目中。在智库的第一个任期结束后，我获得了奖学金，以表彰我对亚历杭德罗·布尔迪西奥（Alejandro Burdisio）的Milqui的3D渲染。我终于觉得自己走在正确的道路上，到目前为止，我在整个旅程中学到的一切都只是让我毫无保留地完全投入3D艺术。
Goals
First, I’m a HUGE fan of Spider-Man, video games and of course Marvel in general. My wife and I spent a lot of time together watching the Marvel movies and rewatching them again in chronological timeline order. It’s a great time to be a Marvel fan and what Insomniac Games is doing with Spider-Man is just thrilling—that game alongside Miles Morales is just insanely fun and made with so much heart. And now we have Spider-Man 2 and Wolverine games to look forward to! It’s just so exciting, I can’t wait!
ScenePriortization-Breakdown
When it came to MJ’s Apartment, I spent a good week planning using PureRef and Photoshop. PureRef to gather all my images, references, and create an asset list. Photoshop helped break the concept down into categories of primary, secondary and tertiary groups—whatever assets fell within those groups were the priority in which I addressed them. For example, the big things like walls, windows, doors, rugs and curtains encompassed a large amount of space in the concept so these became primary assets. Secondary assets were classified as things like furniture—couch, coffee table, bookshelves, end tables, window AC Unit, etc. And tertiary assets were everything that could be considered clutter like books, loose papers, picture frames, etc.
While planning this scene, I kept in mind that I had only 4 months to complete it. This was a project for school after all—so even though I wanted everything to make it into the scene, I clarified some ‘must-haves’ that would take priority over the other details. That way if I had to cut some things for time, at least the story-elements were visible.The kitchen was also a stretch goal. I considered it a secondary asset due to its size, however it isn’t entirely visible in Dennis’ concept. So it could definitely get cut if time is an issue at no cost to the story. But I wanted to try and get it in there to pay homage to the cutscenes we see in Insomniac’s Spider-Man. Watching where Peter stands to wash a dish and how he’s positioned to turn off a burning gas stove, I was able to rough the layout of the kitchen and use additional studio apartment kitchen images as reference.
KitchenLayout
With the concept broken down, I started to identify and label everything that I saw into what later would become my asset list. You’ll see things like “First Pass Done” as a way for me to keep track of what assets need to be created, polished, and finalized. Every asset had at least one or two real life references accompanying them that were similar to what was shown in the concept art. This helped save time later down the road when I started texturing.
MJ_sApartmentPureRef
Overall, my primary goal with this scene was to create a powerful cinematic. It was the first thing that struck me when I came across Dennis Chan’s concept—I imagined the spider suspended on the window sill with the sounds of New York City in the background, the slow camera pan across a cluttered coffee table accompanied by heartfelt piano keys similar to John Paesano’s original score of the video game. It was a flood of inspiration that I clung to throughout the project, which kept me motivated to execute and see it through to the end. It’s been extremely humbling to see the scene and cinematic receive such a warm reception across the game art community.
目标
首先，我是蜘蛛侠，视频游戏，当然还有漫威的超级粉丝。我和妻子花了很多时间一起看漫威电影，并按时间顺序重新观看。现在是成为漫威粉丝的好时机，Insomniac Games与蜘蛛侠的所作所为令人兴奋 - 与迈尔斯·莫拉莱斯（Miles Morales）一起制作的游戏非常有趣，并且非常用心制作。现在我们有蜘蛛侠2和金刚狼游戏值得期待！这太令人兴奋了，我等不及了！
场景优先级-细分
当谈到MJ的公寓时，我花了一个星期的时间使用PureRef和Photoshop进行规划。PureRef 收集我的所有图像、引用并创建资产列表。Photoshop帮助将这个概念分解为主要，次要和第三组的类别 - 无论属于这些组的任何资产都是我处理它们的优先事项。例如，墙壁，窗户，门，地毯和窗帘等大东西在概念中包含了大量的空间，因此这些成为主要资产。次要资产被归类为家具等物品 - 沙发，咖啡桌，书架，茶几，窗户空调装置等。而第三级资产是一切可以被认为是杂乱无章的东西，比如书籍、松散的纸张、相框等。
在计划这个场景时，我一直记得我只有4个月的时间来完成它。毕竟，这是学校的一个项目——所以即使我希望所有东西都能进入现场，我还是澄清了一些"必备品"，这些"必备品"将优先于其他细节。这样，如果我必须为时间剪掉一些东西，至少故事元素是可见的。厨房也是一个延伸目标。由于其规模，我认为它是次要资产，但是在Dennis的概念中并不完全可见。因此，如果时间是一个不花钱的故事问题，它肯定会被削减。但我想试着把它放在那里，以向我们在《失眠症》的《蜘蛛侠》中看到的过场动画致敬。看着彼得站在哪里洗碗，以及他如何定位关闭燃烧的燃气灶，我能够粗略地布置厨房，并使用额外的工作室公寓厨房图像作为参考。
厨房布局
随着这个概念的分解，我开始识别和标记我所看到的一切，后来成为我的资产列表。您将看到诸如"首次通过完成"之类的内容，作为我跟踪需要创建，抛光和最终确定哪些资产的一种方式。每个资产都至少有一两个现实生活中的参考，这些参考与概念图中显示的内容相似。这有助于在以后我开始纹理化时节省时间。
MJ_sApartmentPureRef
总的来说，我对这个场景的主要目标是创造一个强大的电影。当我遇到丹尼斯·陈（Dennis Chan）的概念时，我首先想到的是，我想象着那只蜘蛛悬挂在窗台上，背景是纽约市的声音，缓慢的摄像机平移穿过杂乱的咖啡桌，伴随着发自内心的钢琴键，类似于约翰·帕萨诺（John Paesano）的视频游戏原始配乐。我在整个项目中一直坚持着灵感的洪流，这使我有动力去执行并看到它直到最后。看到这个场景和电影在整个游戏艺术界受到如此热烈的欢迎，我感到非常谦卑。
Modelling
ProjectProgressionGIF
I modeled everything using Autodesk Maya and ZBrush. I created everything from scratch with the exception of the Spider you see on the window sill during the cinematic and the Teddy Bear on the couch—these were assets I obtained for free from TurboSquid which helped save time towards the end of the project.
I was able to quickly rework and texture them to fit the needs of the scene. For everything else, I started with blocking them all out in Maya and bringing them into Unreal Engine one asset at a time. Each asset was shaped very roughly, with no regard for topology or polycount. These block outs are only meant to get a sense of scale and placement to capture the overall composition of the concept. They all are eventually replaced or reworked with a more polished mesh after the block out is completed.
Blockout
After that, I had a good clear picture of the workload in front of me and from then on it was just reworking each and every prop until they were at the quality I wanted. In the case of the laptop and couch, the block out established its size and scale within the scene—I even went further to check the real dimensions of a laptop like the MacBook Pro and couch to make sure they were within reality. So then it was just modeling it out following the references I gathered until it looked right.
LaptopBlockoutFinal
Hard surface props like these were created using beveled edging on the high poly to bake down onto the low poly to get that nice rounded edge. Some props I used separate floating meshes like screws to bake into the low poly as well and utilized face weighted normals by softening edges to take out any unwanted lines/bumps.
CameraBlockoutFinal
Organic props like the sofa required some additional treatment. After the block out, I reshaped it into a mid-poly model that could be sculpted using Zbrush. In the photo below you’ll see from right to left, the initial block out, the mid-poly mesh and the sculpted high-poly mesh created in ZBrush. I created the mid-poly mesh because I wanted there to be enough geometry on it with good topology in order to cleanly sub-divide it while working within ZBrush and also easily optimize it into a low poly to be baked upon. This was a new workflow for me because I don’t often work in ZBrush or on organic models, but I was happy to get to learn ZBrush and look forward to improving my sculpting skills with more practice!
CouchModeling
Books, papers and pictures all had a very simple low poly mesh, with varying sizes but would eventually all share texture atlases so that texturing them individually would be much more efficient. Laying out their UV’s correctly and grouping them together as a single mesh to be brought into Substance Painter was the idea. And it worked out pretty well for me. After the texturing was completed in Painter, I could import all the static meshes into Unreal Engine, create one material for each cluster of books, papers and photos.
造型
ProjectProgressionGIF
我使用Autodesk Maya和ZBrush对所有内容进行了建模。我从头开始创作了一切，除了你在电影拍摄期间在窗台上看到的蜘蛛和沙发上的泰迪熊—— 这些都是我从TurboSquid免费获得的资产，这有助于在项目结束时节省时间。
我能够快速返工和纹理它们，以适应场景的需求。对于其他所有内容，我首先在Maya中将它们全部屏蔽，然后一次将它们带入虚幻引擎。每个资产的形状都非常粗糙，不考虑拓扑或多边形。这些遮挡只是为了获得规模感和位置感，以捕捉概念的整体构成。在砌块完成后，它们最终都会被替换或用更抛光的网格进行返工。
封锁
在那之后，我对眼前的工作量有了清晰的了解，从那时起，它只是重新加工每一个道具，直到它们达到我想要的质量。就笔记本电脑和沙发而言，遮挡块在场景中建立了其大小和规模 - 我甚至进一步检查了MacBook Pro和沙发等笔记本电脑的真实尺寸，以确保它们在现实中。因此，它只是按照我收集的参考资料对其进行建模，直到它看起来正确为止。
笔记本电脑封锁决赛
像这样的硬表面道具是使用高聚乙烯上的斜边制成的，以烘烤到低聚乙烯上以获得漂亮的圆润边缘。一些道具，我使用单独的浮动网格，如螺钉，烘焙到低多边形，并通过软化边缘来利用面加权法线，以消除任何不需要的线条/凸起。
相机封锁决赛
像沙发这样的有机道具需要一些额外的处理。在块出后，我将其重塑为可以使用Zbrush雕刻的中多边形模型。在下面的照片中，您将从右到左看到初始块出，中聚网格和在ZBrush中创建的雕刻高聚乙烯网格。我创建了中多边形网格，因为我希望它上面有足够的几何形状和良好的拓扑结构，以便在ZBrush中工作时干净利落地细分它，并且还可以轻松地将其优化为要烘焙的低多边形。这对我来说是一个新的工作流程，因为我不经常在ZBrush或有机模型上工作，但我很高兴能够学习ZBrush，并期待通过更多的练习来提高我的雕刻技能！
沙发造型
书籍、纸张和图片都有一个非常简单的低聚乙烯网格，大小不一，但最终都会共享纹理图集，以便单独纹理化它们会更有效率。正确布置它们的UV并将它们组合在一起作为单个网格以引入Substance Painter是这个想法。这对我来说效果很好。在Painter中完成纹理处理后，我可以将所有静态网格体导入虚幻引擎，为每个书籍、纸张和照片组创建一个材质。
Books
Texturing
Most of the assets were all treated individually, taken in one at a time into Substance Painter and layered on with materials. First, I bring the asset in and bake in the high poly. Then, typically I start creating folders, masking out the primary parts of the asset and breaking it down to the smallest bits—for example, the camera from the bottom up I broke it down from the base of the camera all the way up to the detailed writing. This just helps to keep my head straight on where everything is so I can iterate quickly if something doesn’t look right.
Substance-Painter-Camera
Before I start layering in my textures, I also set up my Substance Painter with the ACES UE4 Color Profile so that my viewport bears some resemblance to the viewport in Unreal Engine 4.
SubstancePainterSettings
Once all that’s done I get to building up the asset with textures! And then once finished I check it with a PBR Validator to make sure my values aren’t completely out of whack. The PBR Validator is just an approximating tool, so you have to be careful and understand that it’s not always right, but certainly helps if you need to make a judgment call on either the color value or roughness of certain areas of your prop. Green is good, red is not so good!
Substance-Painter-PBR-Validator
The real challenge was figuring out how to texture all the books, magazines, papers, pictures and other small clutter assets. I mean, that’s A LOT of props to texture one at a time! And even more of a performance hit having so many unique shaders all with different textures in a game environment. So I did some research and found my solution by creating prop/texture atlases. This allowed me to group similar assets together, bring them into Substance Painter, texture them, and export one set of textures that would cover multiple assets.
BookAtlases
Having all these books sharing the same UV space allowed me not only to have one set of textures for 10 props, but it also allowed me the freedom to tailor each book with a unique look of roughness, edge wear, etc. I was really excited when I figured out how to do this because it streamlined texturing small props like these and reducing the amount of texture files.
All of the book and magazine graphic covers were obtained for free through Textures.com to save time—I did create The Digital Photography book cover from scratch in Photoshop for fun to add as an easter egg with the Think Tank logo and the name Jason Gullion, my instructor.
纹理
大多数资产都是单独处理的，一次一个地放入Substance Painter中，并与材料分层。首先，我将资产带入并在高多边形中烘烤。然后，通常我开始创建文件夹，屏蔽资产的主要部分并将其分解为最小的位 - 例如，相机从底部向上分解，我从相机的底部一直到详细的写作。这只有助于让我的头脑直视一切，这样我就可以在事情看起来不对劲时快速迭代。
物质-喷绘-相机
在开始对纹理进行分层之前，我还使用ACES UE4 Color Profile设置了我的物质图像库，以便我的视口与虚幻引擎4中的视口有一些相似之处。
SubstancePainterSettings
完成所有操作后，我就可以用纹理构建资源了！然后一旦完成，我用PBR验证器检查它，以确保我的值没有完全不正常。PBR验证器只是一个近似工具，所以你必须小心并理解它并不总是正确的，但如果你需要对道具某些区域的颜色值或粗糙度做出判断，它肯定会有所帮助。绿色很好，红色就没那么好了！
Substance-Painter-PBR-Validator
真正的挑战是弄清楚如何对所有书籍，杂志，论文，图片和其他小杂乱的资产进行纹理处理。我的意思是，这是很多道具一次纹理一个！在游戏环境中，拥有如此多具有不同纹理的独特着色器，甚至更能带来性能打击。所以我做了一些研究，并通过创建道具/纹理图集找到了我的解决方案。这使我能够将类似的资源组合在一起，将它们放入 Substance Painter 中，为它们添加纹理，然后导出一组将覆盖多个资源的纹理。
书籍提篮
所有这些书共享相同的UV空间，使我不仅可以为10个道具提供一组纹理，而且还使我能够自由地为每本书定制粗糙度，边缘磨损等独特外观。当我想出如何做到这一点时，我真的很兴奋，因为它简化了像这样的小道具的纹理，并减少了纹理文件的数量。
为了节省时间，所有书籍和杂志的图形封面都是通过 Textures.com 免费获得的 - 我确实在Photoshop中从头开始创建了数码摄影书籍封面，以有趣地添加为带有智囊团徽标和我的导师Jason Gullion的名字的复活节彩蛋。
Shot_21-scaled
Scattering & Composition
Assembling the final scene and scattering details everywhere was a lot of fun. For the most part, I remained faithful to the original concept. Where I added my own spin was in the pictures having mostly screenshots from Insomniac’s Spider-Man video games with a few free stock photos found online, as well as a couple of other details from the game. Like a Roxxon Power merchandise bag hanging in the kitchen and an Oscorp Industries branded AC Unit.
散射和组成
组装最终的场景并到处散布细节非常有趣。在大多数情况下，我仍然忠于最初的概念。我添加自己的旋转的地方是在图片中，其中大部分截图来自Insomniac的蜘蛛侠视频游戏，并在网上找到了一些免费的库存照片，以及游戏中的其他一些细节。就像挂在厨房里的Roxxon Power商品袋和Oscorp Industries品牌的AC单元一样。
Lighting & Post Production
I used Unreal Engine 4 because of my focus on the game art pipeline. And my lighting setup for both lighting scenarios primarily consisted of a static Directional Light and Skylight as my main light sources, with a customized emissive Sky Sphere HDRI.
First, for the custom Sky Sphere, I dragged in a basic Sphere from UE4 and scaled it up to about 20000 units so it would encapsulate the entire scene. I then created a material for the Sky Sphere that would allow me to add in an HDRI and give me control over the brightness and rotation of the HDRI within the scene.
CustomSkySphere
I then created a material instance of this and applied that instance to my scaled-up sphere and searched HDRIHaven.com to find an HDRI that contributed best to the scene and matched up well with the concept.
With my custom HDRI sky set, I then researched some ArchViz lighting setups and found Luoshuangs GPU Lightmass (). I believe something like this has been integrated already into Unreal Engine 4.27, but I was working under 4.26 during the creation of this project. The documentation can provide more in-depth detail, but essentially this allowed me to achieve much faster light bakes for faster iteration and also at a much higher quality with the settings cranked. The only thing I had to adjust was the Number of Indirect Light Bounces within World Settings and edit the BaseLightmass.ini Primary and Secondary GI samples to increase the quality of the light bakes.
I know that last bit probably sounds very confusing—but if you’re still working under 4.26 it’s fun to play around with. Look up some videos on Youtube and check it out! I honestly tried it just to see what it did and was blown away by the results and how fast it baked. I’m not a fan of baked lighting because of how time-consuming it can become when iterating, so this really helped ease that pain a little. It’s always worth exploring tools like these just to pick them apart to see what they can do—just remember and always back up your project! Safety first kids!
I like to keep my Post Processing settings very simple and subtle. Only making dramatic changes if it favors matching the concept. So under the lens you’ll see my settings in this image. Very low Bloom and Grain settings and modest Chromatic Aberration value. The Vignette felt like it darkened my scene so I adjusted that to be less dramatic. The Exposure settings were adjusted to match what felt appropriate realistically as well as to what I felt was happening in the concept.
PPSettings_01
Color Grading I also barely touched. These settings I feel can be more appropriately managed in something like DaVinci Resolve or After Effects if you are shooting a cinematic, but in general I tend to only make the tiniest adjustments here.
PPSettings_02
Heading further down the Post Processing volume under Film, I usually start the Slope at .8 and make minor tweaks before finding the result that feels right. Toe usually stays at .55 in most cases and Shoulder I take down to .25. Things like Ambient Occlusion get cranked up for the final renders because it’s always good to get nice-looking AO.
Normally I don’t touch the Global Illumination settings, but I had to adjust the Indirect Lighting Color to get a cooler greenish-blue feel like what’s in the concept. Without it, things looked way too warm. I turned on Motion Blur because of the cinematic I planned on shooting, set the Reflections to Screen Space and cranked up the quality.
PPSettings_03
Below are two screenshots, first with the Post Processing Volume On, and the second with the Post Processing Volume Off.
灯光和后期制作
我之所以使用虚幻引擎4，是因为我专注于游戏美术流程。我在两种照明场景中的照明设置主要由静态定向光和天窗作为我的主要光源组成，并带有定制的发光天空球HDRI。
首先，对于自定义的天空球体，我从UE4中拖动了一个基本的球体，并将其放大到大约20000个单位，以便它将封装整个场景。然后，我为天空球体创建了一个材质，允许我添加HDRI，并让我控制场景中HDRI的亮度和旋转。
自定义天空球体
然后，我创建了一个材质实例，并将该实例应用于我的放大球体，并搜索 HDRIHaven.com 以找到对场景贡献最大且与概念非常匹配的HDRI。
使用我的自定义HDRI天空集，我随后研究了一些ArchViz照明设置，并发现了罗双GPU Lightmass（）。我相信这样的东西已经集成到虚幻引擎4.27中，但是在创建这个项目时，我是在4.26下工作的。文档可以提供更深入的细节，但从本质上讲，这让我能够实现更快的光烘焙，从而实现更快的迭代，并且在设置启动的情况下也能获得更高的质量。我唯一需要调整的是"世界设置"中的间接光反射数，并编辑了"基本光质量.ini主要和次要 GI 样本，以提高光照烘焙的质量。
我知道最后一点可能听起来很混乱 - 但如果你仍然在4.26下工作，那么玩起来很有趣。在Youtube上查找一些视频并查看一下！老实说，我尝试它只是为了看看它做了什么，并被结果和烘烤的速度所震撼。我不喜欢烘焙照明，因为它在迭代时会变得多么耗时，所以这真的有助于减轻一点痛苦。探索这样的工具总是值得的，只是为了把它们分开看看它们能做什么 - 只要记住并始终备份你的项目！安全第一的孩子！
我喜欢保持我的后期处理设置非常简单和微妙。只有当它有利于匹配概念时，才进行戏剧性的改变。所以在镜头下，你会看到我在这张照片中的设置。非常低的绽放和颗粒设置以及适度的色差值。小插图感觉它让我的场景变暗了，所以我调整了它，使其不那么戏剧化。对曝光设置进行了调整，以匹配感觉合适的现实情况以及我感觉在概念中发生的事情。
PPSettings_01
颜色分级我也几乎没有碰过。如果你正在拍摄电影，我觉得这些设置可以在DaVinci Resolve或After Effects之类的东西中得到更合适的管理，但总的来说，我倾向于只在这里进行最微小的调整。
PPSettings_02
在"胶片"下的"后期处理"卷下，我通常从0.8开始"坡度"，并在找到感觉正确的结果之前进行一些小的调整。在大多数情况下，脚趾通常保持在0.55，而肩膀我则下降到0.25。像环境光遮蔽这样的东西会在最终渲染中被调高，因为获得漂亮的AO总是好的。
通常情况下，我不会触摸全局照明设置，但我必须调整间接照明颜色，以获得更凉爽的绿蓝色感觉，就像概念中的内容一样。没有它，事情看起来太温暖了。由于我计划拍摄的电影效果，我打开了运动模糊，将反射设置为屏幕空间并提高了质量。
Conclusion
The scene was completed in about 4 months—starting from October 4th, finishing around January 23rd. I think the biggest challenge was just believing if I could pull it off or not. I have always been pretty good at managing my time and finding balance, there weren’t many late nights working on this, although I did work full-time every day on it. I took breaks on the weekends, but sometimes I’d lose a weekend or two if I got stuck on a problem or had to play a little catch-up. But yeah, I just mostly struggled with the idea of if this piece would come out good or not and if I could pull off everything that I envisioned in my head. I think I was successful though.
I’m always looking to improve. To be honest, I’m already preparing my next environment piece, but this time I’ll be venturing into Unreal Engine 5 to explore what possibilities are there. I’ve been learning a lot about Nanite and Lumen and would like to incorporate that into my workflow and see where things go. I don’t know where any of this work will lead me, but I’m always moving forward. Thanks so much for giving me this opportunity. As a student, I don’t know if I really educated anyone, but I do hope I at least inspired someone to give this 3D Art a try. You never know what you just might create!
结论
场景在大约4个月内完成 - 从10月4日开始，到1月23日左右完成。我认为最大的挑战就是相信我是否能成功。我一直很擅长管理自己的时间和找到平衡，尽管我每天都全职工作，但工作时间并不多。我在周末休息，但有时如果我遇到问题或不得不追赶一两个周末，我会失去一两个周末。但是，是的，我只是主要纠结于这个想法，即这件作品是否会好，以及我是否可以完成我在脑海中设想的一切。不过，我认为我是成功的。
我一直在寻求改进。老实说，我已经在准备我的下一个环境作品，但这次我将冒险进入虚幻引擎5，探索其中的可能性。我一直在学习很多关于Nanite和Lumen的知识，并希望将其融入我的工作流程中，看看事情的发展方向。我不知道这项工作会把我引向何方，但我一直在前进。非常感谢你给我这个机会。作为一名学生，我不知道我是否真的教育过任何人，但我确实希望我至少能激励某人尝试一下这种3D艺术。你永远不知道你可能会创造什么！
|
