Who do you design for?

Hello,

I'm not sure this is the right forum, but it seems like the best choice for this question.

I am building things in SL and obviously I end up using textures on them. My question is this: Knowing that residents have all different kinds of graphic settings and video cards, what level do you design for?

Things can look so different based on settings and graphic ability of the viewer's computer. When I look at something I make using "shiny" on the textures, it looks really nice and the look I want. However, without shiny it looks totally different. So -- what to do? Do I use Shiny knowing that not everyone will see it that way? Do I photograph the item in world with the shiny on and hope that the person buying it will see it that way?

I have bought many things from pictures and then the item didn't look quite the same in person - obviously due to lighting, graphic settings, etc.

I want to build things that look as good as possible obviously, but also don't want to dissapoint people that may see them differently.

Does anyone have any opinions on this? I'd appreciate any feedback.

Thanks,

Sparkly

Personally, all my items are made for me and how I view them. I make them for a specific purpose...usually some form of RP. If I think them good enough, then I will post them for sale. Not sure how you would be able to please everyone; thus, I try to please myself. If a customer has an issue, I will try to accommodate them if possible.

I'll usually build assuming most of the standard client features will work for the people experiencing them. Local light, shiny, and the like. I'll usually check a build with some of those turned off, though, so I know it will still look 'acceptable' if they don't have Shiny turned on. And for example, something that casts local light will also use at least some fullbright, so they see something "glowing". Haven't mande anything yet that really makes good use of "glow", but usually if that is off, most people who haven't been seeing glow before won't notice its absence, either.

I would always assume that basic shaders and graphics set for prims to be high, as those are the two most extremely basic requirements for viewing and using SL properly in my opinion as otherwise everything will look heavily degraded and out of touch as everything is made with these two settings in mind to be equally throughout SL. Local lights not so much as it lags and is not noticeable as an effect.

Im probably seen ignorant here and appealing towards the graphics appetisers, but I would like to note I am on a crappy system and the objects detail is the only slider on my graphics setting set on high with everything else set to low so that all the objects I see are mathematically/geometrically accurately represented as they are to be seen as they were made to, but FPS is kept reasonable for my low specs.

I would recommend using all the non-lag graphics options available as would be used on average by your ideal consumers, rather than the normalised population. Shiny and glow have negilible effects on FPS compared in contrast to their advantages of providing a more dynamic life on objects. I saw once a horse furry avatar on sale at Slexchange whereby the creator took the picture when they had their objects detail set to extremely low, the horse ended up being starved on polygons and looking like some mount from World of Lowpoly. Not many sales reviews on that one and looked like it ended up having really bad ranking .

Great question. My best answer is play to your audience. Think about who will be your most likely user base, and what kinds of computers are they likely to have. It's all about balancing three key factors: visual quality, performance, and usability.

All projects need an adequate amount of each. Skimp on any one, and your work will be bad. What separates the great artists in SL from the OK ones, in my opinion, is not just that they include all three factors, but how they balance them.


Here are some general examples of what I mean. Note, I'm writing from the perspective of doing contract work, since that's mostly what I do in SL, so not every last detail of this will be directly applicable to what you might be doing. But the general principles are applicable to every project in all of SL.


1. If the project is for a stereotypical business client, my primary focus will usually be usability first, performance second, and visual quality third. That's not to say visual quality isn't still extremely important. It absolutely is. It's just not as high on the priority list as the other two factors. For business clients, the design goal is to make it so business grade laptops, and business-minded users, will be able to utilize the finished project successfully.

I'll talk about the user needs first. Forgive me if I over-generalize a little in what I'm about to say. I don't mean to dehumanize anyone by lumping people into categories. This just happens to be the simplest way to explain my point in this context. Obviously, there is room for exception.

Business clients in SL are an interesting breed. Although they could probably kick my ass every day of the week when it comes to filling in spreadsheets at speed, most aren't usually very experienced when it comes to things like basic navigation of an avatar, POV controls, etc. There are exceptions, of course, but in general, business people aren't usually coming from a gaming background, so this sort of thing is brand new to them in many ways.

My job, therefore, is to go to pains to make it as easy on them as possible. I'll make sure buildings are super easy to navigate, no unnecessary twists & turns, plenty of roof access, etc. Any signs will be easy to understand. Interactive items like teleporters and such will always be clearly marked "Click here to do _________". The goal is to have as little margin for error as possible.

Now let's talk technical needs. There's no easy way to say this. Designing for business grade laptops sucks. They tend to have either integrated video, or ultra low end discrete video. Neither is well suited for SL. Also, they tend to have either not a whole lot of RAM, very slow RAM, or both. On top of all that, they also tend to have fairly crappy CPU's. All in all, business machines are not a good fit for SL.

But if business machines are what the client has, that's what I have to make work. Here's what I'll typically do. I'll keep textures as few and small as possible, painting only what is absolutely necessary to get the client's message across. I'll try to stick with geometry that is relatively easy to render, and that will be recognizable even at low LOD. I'll limit the amount of transparency in use. I'll usually avoid shine, local lighting, glow, particles, etc.

That may sound like a recipe for ugly, but it doesn't have to be. It's entirely possible to create a visually pleasing build while keeping it simple.



2. If the client is a TV network, movie studio, or some other media entity, then the balance is exactly the opposite of the business build. It's visual quality first, performance second, usability third. Again, that's not to say that usability isn't still important. It's just that the visuals are the main point of a film-quality build, so that's what has to come first.

For a media industry project, it's sculpties, sculpties, and more sculpties, full lighting bakes, large high-detail textures, unrestricted use of effects, etc.



3. For more "miscellaneous" kinds of clients, like universities, government agencies, retailers, etc., whose users and their computers will be a very mixed bag, I try to find the best happy medium I can. The three factors weigh more or less evenly, wherever possible. Texture sizes and amounts are not exactly restricted, but they're kept reasonable. Effects like glow and shine are used sparingly.

The focus is on how to appeal to the broadest possible audience. Some low end machines will end up left out, but there's not really any getting around that if you want everyone else to have the most compelling experience possible, given the likely "average" level of experience with SL and with video games.



I think for the majority of projects in SL, number 3 is the rule of thumb. 1 and 2 are more specialized cases. 3 represents the masses.

I guess you could summarize it like this. Design for the widest possible audience, with what strikes you as the best achievable balance between visual quality, system performance, and human usability, for each project you do. Don't be afraid to leave out some low end users if you have to, but try to be as inclusive as possible.

Here are some general tips to think about each time:

1. Is that 512x512 texture absolutely necessary? A 256x256 will load 4 times faster, and use 75% less memory. Will a 256x256 do the job?

2. Are post-render effects like shine and glow absolutely necessary? Can you incorporate a similar look into your texturing, so that low end users can see it? Or better yet, can you use both?

3. Are you benefitting at all from that local light you just placed on your front porch? Remember, as soon as some jackass with a half a dozen face lights shows up, that porch light is going to turn off. Could you acheive the same (or better) look by baking lighting into your texturing instead of using an actual light? (You'll find the answer to this one is almost always a resounding yes.)

4. Are you using transparency only where absolutely necessary?

5. Are you balancing your poly counts and your prim counts appropriately. Remember, a single sculpty can make two or three really nice couch cushiouns, to take the place of a dozen or more prims, which seems like a great trade. But when you consider that that single sculpty might have 10 times the poly count of the regular prims, the performance hit might not be worth the prim savings. Weigh your options carefully.


I think you're getting the idea by now. There's plenty more that could be said on this subject, but I'll leave it at that.

I can't stand shiny on "organic" type sculpts. It looks hideous in my opinion. The only exception to this is faceted surfaces such as gems. I think that making sculpts shiny is sort of an attempt to hide poor textures.

I don't design for anyone. What I try to do is make stuff that is the most efficient possible for a desired level of detail - sculpt wise and texture wise. The principle is that this will allow either for extremely low rendercost builds, _or_ more complex builds with more visual detail.

I bought a fair amount of stuff before I was able to get Shiny working on my computer, and afterwards I decided a lot of it looked better NOT shiny. Some was no-mod, so not only had I looked worse than I thought all that time I couldn't fix it. That really put me off shiny for a long time.