在网上看到了一篇 Sarah B. Nelson 写的文章:

No designer is an island

觉得写的很好 所以摘抄了下来

特别对于像我们学design的人

看了后受益匪浅

--------------------------------------

When I suggest collaborative design to some designers, I often hear, “Yuck! Collaboration is just design by committee!” They aren’t wrong. Poorly facilitated collaboration can kill a design project and demoralize a team. But the truth is, you can’t not collaborate. Whether you like it or not, you have to work with other people. You can work at them or with them. It’s your choice.

As designers, we exist in a complex ecosystem, balancing the needs of diverse constituents, striving to do the best work we can. When approached strategically, collaboration can help you communicate more effectively with stakeholders, gracefully negotiate political waters, and use collective experience to generate ideas or inspiration. Armed with rich, first-hand source material, you can confidently make informed, intelligent decisions.

I advocate for a very specific type of collaboration; I call it structured collaboration. It’s not rocket science but it can be a powerful addition to your toolbox. Unlike simply going into a room and working together informally, structured collaboration consists of thoughtfully designed work sessions, using visually-based techniques, physical materials (stickies, paper, pens) and planned activities to move the design process forward. Structured collaboration is loosely based on Participatory Design techniques, where users are directly engaged in the design process. Since these users are typically not designers, games, activities, and other visual tools help them express their ideas in ways designers can interpret. Structured collaboration does not replace other design tools; it simply helps designers better understand and balance the needs of a diverse set of constituents.

As designers (in the broadest sense), we focus a lot of energy on technical skills but often neglect or downplay the softer “people” skills. Daniel Goleman calls these skills “Emotional Intelligence.” They are as important to our success as IQ and technical proficiency. Qualities of emotional intelligence include, amongst others, social skills and empathy. Social skills, in Goleman’s definition, include “proficiency in managing relationships and building networks” and “an ability to find common ground and rapport.” If you possess social skills, you will be more effective in leading change and being persuasive.  Empathy includes “the ability to understand the emotional makeup of other people” and “skills in treating people according to their emotional reactions.” We talk about empathy with our users but we also need to cultivate empathy for our stakeholders, too.

Think this sounds fluffy? Yeah – I used think the same way. I’ll make it more concrete. Imagine this typical design process for a home page: you do some stakeholder interviews, you write up a brief, your team goes away, works hard, and spends weeks developing some brilliant concepts. You give your big presentation to your key stakeholders and BAM! You’ve totally missed the mark, or at least your stakeholders think so. Everyone is disappointed and frustrated.

How did that happen? You’ve failed to communicate, to find common ground and build trust. Assuming your design skills are rock solid and your stakeholders aren’t evil (and they almost never are), you’ve missed some critical information. This brings us back to social skills and empathy, right? For instance. consider the possibility that your stakeholder hadn’t completely thought through the problem until you showed him something. After all, he’s not the designer. You are. Or, what if he is worried about his own department’s performance? Perhaps, since he lacks a design vocabulary, he is asserting his needs the only way he knows how. His reaction might be more emotionally charged than normal. He’s had few other outlets up until this point, so he’s using the presentation to let you know his concerns. There could be a hundred other reasons that you are unaware of.

When you take a structured approach to collaboration, you can nip some of these communication problems in the bud. You allow people a chance to weigh in. You establish communication lines and begin to better understand how people work, what results they need, and how to best play to their strengths. By holding carefully planned work sessions, you can use visual techniques and activities to answer questions, foster discussion, and bring your whole team — stakeholders and all — together around a specific problem.

How would this work? In the design problem I mentioned above, you might first hold a two-hour visioning work session with your stakeholders. Together, you’ll imagine the possibilities for the home page. You might ask them to draw their ideal home page or assemble it from pre-made parts. You might ask them to share examples from other sites that work for them. You might bring in your own examples. The examples or drawings are not the interesting part. What they say about them is. Listen closely for concerns, preferences, and preconceptions. You don’t necessarily have to accommodate them, just be aware of them. This is your raw material.

Half-way through the project, spend an hour with the same stakeholders to look at, discuss, and evolve some early concepts. Hang your concepts on the wall. Ask for specific kinds of feedback and post questions you’d like answered on the wall next to the concepts. Point out concerns you have. Recall the concerns you heard expressed in the vision session and show how you’ve addressed them. Give people sticky notes and ask them to write down their concerns or questions directly on your concepts. Again, you don’t have to accommodate all the feedback, just take it into consideration. Your goal is to hear their concerns and needs early while you can still address them. They can also hear your concerns early and take them into consideration.

The real benefit of these two simple sessions is trust. You are building a common ground and sense of rapport. Finally, when you share your work, your stakeholders won’t be surprised and they are less likely to nitpick it; they’ll have seen ideas evolve from the beginning and better understand how to engage with you. They also know that you’ve heard their concerns and have addressed them. Your critique will be more productive and informed critique. 

One word of caution: you are still the designer and still need to do the design work. Don’t get run over. Work sessions are made or broken on effective facilitation. This means that, as the session’s facilitator, you need to set the stage for the session, elucidate expectations, and handle conflict gracefully when it arises. If you don’t think you can facilitate a session alone, get help from a manager, project sponsor, or another designer. Whatever you do, don’t abdicate authority to the loudest mouth in the room. Be polite, but assertive. If the session goes south, table it. Take a break. Re-group. Be clear about what you need and what you can offer. Ask stakeholders what they need from you and be clear about what you can and can’t do. If you are critiquing design work, just critique it. Don’t solve it on the spot. If you are generating ideas, generate ideas. Don’t commit to one.

Build collaboration and work sessions into your practice. It might be daunting at first but try it. Start small and simple, gradually including your stakeholders. Be patient through the transition; there are bound to be fears and missteps. Keep trying to foster strong communication and clear expectations through work sessions. Over time, you’ll get better at it and people will trust the process more. You’ll be able to make decisions more confidently. The social relationships and empathy you develop will not only make your job more satisfying but will help you design more efficiently, and increase the quality of your work.

1. Create a base mesh in Maya or 3ds Max. As quick as possible, with a subdivision friendly topology. Usually following the skeleton rig that was provided to me.
2. When I build my base mesh, I keep track of the amount of polygons it has. And since I know the limit that my system can handle (without HD mode) in ZBrush, I make sure that I can get close to this amount after subdividing my mesh. For example, my office computer can handle a 6 million polygon SubTool. So I made a base mesh that had ~6000 triangles. A mesh with 1500 triangles would have worked too; I would just need to subdivide one more time to reach the limit. It is a bit of a math job, but it is definitely worth it. The last task I had was a full body, including the head. And if I got stuck at 1.8 million there are a lot of things I would not be able to bring to a fairly detailed level. On this task, for example, I also added a single extrusion at each toenail and fingernail to make sure that, once subdivided, I would have enough geometry in those areas to get sharp-looking nails. So I put in a lot of thought during this pre-ZBrush step. Preparing carefully can save you a lot of time in the end.
3. Once I jump in ZBrush, from the simple base mesh I will sculpt all my major masses, main muscle groups, study silhouette, and so on. I do not sculpt any details at this point.
4. Once I got the major masses locked down, I will work my way up, one subdivision level at a time, until I get all my details in. I shift constantly between different Subdiv levels.
5. I will use 3D layers for asymmetry, skin pores, veins, and all types of details that can benefit from being turned on/off.
6. I will also use layers to make various iterations of my sculpt, like using different muscle volumes, muscle definition, style and so on. I will do that to give a few more options to my art director. Those variations are quick to produce and are usually very welcomed by the decision makers.
7. Once the high-res sculpt is approved, I will create a 4K texture for my high-res model. I will send my lowest subdiv level to Maya (and/or Unfold 3D) and do a quick UV job on it. The important thing here is to have as little distortion as possible in the UV's. This will not be the UV's I will use for my final game mesh, so it is free of restrictions at this point. The reason why I like to texture my high-res object first is to have a solid “center point” to work the game mesh around. There is always a lot of surface sampling, mesh transferring, mesh modifications, etc. happening during a production, so it is good to have a high-res, undistorted source texture to work from. Once the texture is done, I get an approval on it. I will sometimes go back to the sculpt and make some minor tweaks (like sharpening some plane changes) based on how it looks with a texture.
8. I will build a game mesh, using the high-res mesh as a guide. I like to use 3dsMax and the PolyBoost plugin to do so. It allows me to focus solely on topology flow and triangle count, since I draw my polygons straight on the surface of the high-res mesh. So no need to worry about form.
9. Once the game mesh is done, following all technical requirements, I will create my UV's in Maya (and Unfold 3D). This UV set will be packed pretty clean and tight, trying to keep straight borders around my UV shells, so that it works fine with different L.O.D. meshes. Using the Unfold option while keeping the UV borders fixed in Maya is extremely useful to get rid of UV distortion quickly.
10. Once this is done, I just need to do a surface sample pass. Projecting Normals, Diffuse and Ambient Occlusion maps to my game mesh. I do so with xNormal. Very fast and efficient.
11. Once the game mesh is ready, along with its texture and a high-resolution mesh is done with its high-resolution maps as a source, I hand off the model to a setup artist. At this point, there might be some back and forth to adjust the game mesh, depending on the needs.

What are your preferred techniques for creating that initial high resolution mesh?
I like to work one subdivision level at a time. It gives me a better control over the forms this way. I make sure that I focus on the main masses at first and get into more refined and smaller shapes later on. I am not going to say that this is the one-and-only way to do things properly, but this is the way I like.

Once I reach my highest level I will start to work details, often using layers, in order to have the options of toggling them on and off if I need to.

While sculpting, I have shortcuts setup for the brushes I mainly use: Standard (playing with BrushMod very often), Inflat, Pinch, Flatten, Magnify, Move and Clay. These shortcuts are a huge time saver, since I cycle a lot through different brushes when I work. I also like to get rid of any unnecessary interruptions, even if little, that would break my creative flow.

How about the chain mail? Uniformly repeating effects like that and Baraka's leathers are something that people often ask about.
Those can be tricky. It is always necessary to take a step back before tackling something like this. I spent a fair amount of time trying to figure out what was the most optimized way of doing so.

In the case of Baraka’s scales on his armor, I tried a quick Hair solution in Maya. Using scales as my hair strand, letting gravity pull them down, in a natural state, resting on each other. Failed! But at least I tried.

Since my ZTool (with all the SubTools) was getting pretty heavy (around 10 million polygons), I decided to model the scales in Maya and smooth them out in ZBrush. I could have tried to sculpt them by hand in ZBrush, but it would have required more time and a lot more polygons.

For Raiden’s chainmail, I created a ring pattern in Maya, making each ring fatter so that the subdivisions in ZBrush would bring them to a regular size.

So, in Maya, I placed this ring pattern around the spots where Raiden needed it. I deleted all the rings that were bleeding out of the armor area. I used deformers to adjust it a bit to the curves of his body. Then, In ZBrush, I imported the “low-res” chainmail as a SubTool and subdivided it twice, to get smooth-looking rings. I also used the Move tool to reshape it better to the forms of Raiden’s body, taking good care that the rings would not stretch.

The final ZTool stands at 20 million polygons. I am quite amazed at how great ZBrush is in handling such a huge amount of information. It was still workable, even with a 32-bit system that is far from being top-of-the-line.

What is the process that you use for texturing your models?
I like to sculpt in the details as much as time allows. Then I will use a Color map to get that extra touch of randomness that can make a surface look more natural. In the case of a character’s skin, I will use ZAppLink extensively. I will project textures taken from Midway’s model library or 3d.sk and align everything with my high-resolution mesh, not the other way around. I will remove any shadow and highlight information from the skin, keeping only the color variations and some smaller information like wrinkles and skin bumps, if needed. But, as I said, I usually prefer to sculpt those because I get more control over the final result this way.

Once the projection phase is done, I will do all the tonal adjustments in Photoshop and so on and so forth. Then, back to ZBrush. ZAppLink is an essential tool for fixing seams. I pay good attention to fixing all of them, especially in areas where it would be easy to say: “nobody is ever going to see that” like behind the ears and around the fingers. How many times do you end up in a cut scene, with a camera straight behind your dude’s head with an extra large view of his ears? Or another guy points his finger straight at the camera? Seams in these cases would be deadly. So that is why I love ZAppLink so much.

I also spend a lot of time working on the specular map. I consider this map to be as important as the normal map. This is the texture that will make your materials read the way they should. A few tricks can help fake some nice effects, like reflected light. For example, if my character is wearing a golden piece of armor on his forearm I will paint a yellowish color on the skin, right at the base of the armor in the Specular map. When the light hits that area, it will give the illusion of reflected light. A small detail that might be consciously unnoticeable, but it makes a subconscious difference in the overall result.

I try to create my source textures as high resolution as I can, depending on the time I have. Otherwise you often end up in a situation where a texture created at 1024x1024 doesn’t hold up very well in a cinematic close-up shot.