From: Rex Tardis
I have a question as to what is the best way to handle a large structure like a skyscraper's texture.
The structure is 20 meters wide and 30 meters tall using 6 10m X 10m prims. I have a texture that I've sized to 1024x512.
Is it better to:
1. Cut the image up to 6 512x512 textures and post one on each prim?
Or
2. Use the whole texture on each prim and adjust the size and offset of the position.
My question is based on if you use the same texture on multiple prims, only one texture download is done for all prims as opposed to (in this case) 6 separate texture downloads.
I want to make sure I can cut down on lag and expedite the loading of the textures.
Thanks!
Your second option is the way to go. You are correct that one download is more efficient than six separate ones. Don't cut up your texture.
As White said, splitting over 3 parts won't be mathematically perfect, but it will be close enough that you won't notice. If you want to ensure there's not a single lost pixel though, you could pad your texture on a larger canvas. That might be the best thing to do anyway since your 3:2 height to width ratio doesn't fit neatly into power-of-two sizing.
Here's what I'd probably do. Start with a texture size that is right for the proportions of your build, say, 256x768. Then mount that texture on a 256x1024 canvas. When you apply the texture to you build, the repeats will end up needing to be .5 on the horizontal, and .25 on the vertical. Offset accordingly to show the entire image.
You're wasting 1/4 of the canvas as empty space this way, but you could recover that by using that quarter for another texture to go elsewhere in the build, and then nothing will have been lost. That's what I'd do if it were me.
From: Abraham Attenborough
what file-formats are allowed to use?
maybe u can use vector graphics like these?
Abraham, the file formats that SL can use are listed in the sticky at the top of this forum, entitled "Texture Size, Pixel Counts, Video Memory, and File Formats". I'd highly recommend you read it.
So you know, those PSD's on Robin's website are stictly for use on your own computer when making clothing. SL cannot use them directly.
Also, just for clarity, they're not actually vector formats. They do contain what Photoshop calls "vector smart objects", but they're still raster images. While vector smart objects do use some vector math, which is where they get their name, they're applying that math to a raster source object to spit out a scaled raster target object. It's Photoshop's way of making the resizing process a bit more lossless than it otherwise would be. It's not producing an actual vector image file.
Even if it were though, I don't see how that would be of any particular benefit for spreading a texture over multiple prims. Care to explain your reasoning for why it would?