A little story on wasted time. I've recently upgraded the frame rates of the spinning TriOp and Med/Sci logos (again), and as a result I now have
68 256x256 textures to deal with. So I've been looking into saving them as DDS to save video card memory. The advantage of DDS files is that they support DXT compression. DXT for opaque textures yields a fixed compression ratio of 8:1, and it can be decoded directly on the video card in realtime, which saves a ton of the aforementioned VRAM.
However...
As it turns out, the difference between 68 textures saved as PNG vs DDS is 23 megabytes of VRAM. That may sound like a lot, but if you have, for example, a bargain-basement 4GB video card, 23 MB is barely half of one percent of that. SS2 as a whole, with SHTUP and 400 enabled, barely breaks 700MB total VRAM usage.
So in the end, not worth the hassle. All that VRAM exists to be used, after all. And it's a small price to pay for not having to deal with compression artifact nonsense like this:
https://www.youtube.com/watch?v=0TtH4ldEzXAFor those wondering what they're looking at here, DXT compression works by dividing the image into 4x4 blocks, then within each block directly encoding the TWO most prominent colors, then deriving two more colors interpolated between those colors. It works great for photographic images, but it absolutely sucks for images with fine multicolored detail. Like for example, an image where green, orange, and black might all be present within the same 4x4 block.