3d object rotation again.

I’m coming back to an old topic, because the question has changed a little bit.

I have this sprite, it’s a very simple sprite, but graphic has been pulled from a Blender model. It’s a simple cylinder shape with a bit of detail, kind of like an oil barrel I guess.

Now, I was looking at a way to rotate (tumble) the Blender object, then render it to a target and display the resultant graphic as a sprite. I was dissuaded from doing this as it may be too resource hungry.

So the obvious answer is to create a sprite sheet. I bought a bit of software for creating sheets which works really well, but the results were less than pleasing. To get the smoothness of animation I really want I would need hundreds of graphics. This itself isn’t a problem, but I started getting warnings about sprite sheets usually being no larger than xxxx,xxxx on most devices, something like 2800,2800 or something.

The question now is, how many sprite sheets can you stuff into memory, and what’s the most efficient way of doing it? I have to admit I don’t know a huge amount about graphics cards, I know graphics are loaded into the card and these obviously need memory, but I have no idea how they do it, is it byte per byte, or is there any kind of compression, is there a height/width of memory canvas, are there layers, does it matter what G card you’re talking about, this phone, that tablet, etc, etc!

Or, is there another way? Someone suggested I write the interface using a camera in orthographic view. I would need to completely re-write most of my code, but I guess I would get the look I want. Would I have to change all my sprites to objects, or can you mix and match sprite batches with a camera?

Any and all musings welcome!

Thanks.

MuntyScruntFundle, I’m very interested in this problem as I wanted to do something similar in MODO 7.0 - set up a simple object, rotate it around a single axis, and save to disk a number of images so that later I could add them to a spritesheet.
I wanted to set up my camera in MODO in such a way that only what was inside the object’s 2D bounding box was rendered to disk, but I never figured out how to do it. Have you any idea how that might be done in Blender? Thanks.

Yeah, Blender should do the trick for you, as long as the scene isn’t complicated. I ended up rotating my camera around the object and simultaneously revolving the object on a different axis, it gives a pleasing effect, certainly good enough. Blender has a transparent ‘film’ setting so you don’t get any background rendered.

Give it a go, I’m getting to be ‘ok’ with Blender now, if you have any questions please don’t hesitate to ask.

Cheers.