FreeOrion

Geoff, how does this affect performance on your laptop?

tzlaine wrote:Geoff, how does this affect performance on your laptop?

Assuming you mean changing the size of the gas textures, with a 500 star galaxy, at full zoom out, with gas textures 16 times the system icon size, I get 7 FPS. With gas textures 12 times the system icon size, I get 9 FPS. (Exact FPS may vary with what else I've got running and the phase of the moon, but these tests were in similar conditons.)

Can we get these two related things added in before the next release?
* each nebula appears only once per galaxy.
* nebulae are rendered above all gaseous substance sprites (the gaseous stuff can obscure the special nebulae)

and maybe also:
* nebulae are located only centered at the coordinates of a star. (simplest way to prevent nebulae from appearing in weird locations)

As well as:
* add an option to turn background nebulae on/off

v0.3.9 was actually put up last night, though I've been too busy to make a post about it. Those changes might get in the next version, but it probably won't be as soon as you might have expected...

Geoff the Medio wrote:Those changes might get in the next version, but it probably won't be as soon as you might have expected...

From my outside perspective, they seem like a a good project for someone relatively new to FO code to try their hand at.

Update: The rendering order and toggle have been committed.

Thanks for the new release. Just FYI, I don't have any performance decrease. I get 59-60FPS, as I always had.

This might be a stupid question, but where are the nebulas in version 0.3.10? Is their absence a bug or something that was done deliberately? Or is it just me not noticing some option to turn them on/off?

I wasn't aware nebulae were removed, and I don't know of any added toggles related to them. Try using a bigger galaxy; you might just not be seeing any nebulae at a small galaxy size...? If they're entirely gone, then that's a bug.

Geoff the Medio wrote:I wasn't aware nebulae were removed, and I don't know of any added toggles related to them. Try using a bigger galaxy; you might just not be seeing any nebulae at a small galaxy size...? If they're entirely gone, then that's a bug.

There is only the gaseous substance and no "special" nebulas. Not even in a 500 star galaxy.

I had temporary deactivated them, by using a single empty nebulae image and moving the rest to a sub folder. Until there is some way to control the amount and color of nebulae and there is some actual game play effect caused by nebulae, I don't want them to uglify the galaxy.

Further thoughts on this:

- Nebulae should be categorized by dominating color. This can be done similar to the gaseous texture sets, by putting similar nebulae in the same folder( one for orange nebulae, red nebulae, cyan nebulae, etc). We can then even include color variations of the nebulae we already have.
- Certain Nebulae can be combined with certain gaseous texture sets. This requires some art direction and should not be done at random. "Galaxy styles" could be defined in a .txt or .xml file and then chosen at galaxy creation. Such a file would list gaseous texture sets and then possible nebulae assignments for those.
As mentioned previously:
- Nebulae should not be used more than once per galaxy, even if there are variations. This can easily be solved with the proposal above.
- Nebulae should get flipped and rotated arbitrary.

The engine that makes the gaseous substance that gives better weight to the galaxy is pretty cool. But i was poking around to see how it works, and i caught some glitchy stuff. Obviously not an urgent problem, but something that's worth noting.

The following example uses this test texture: The result in a small galaxy is this: Issues:

• * Any pixels on the outside edge of a gaseous graphic get stretched a long way. Doesn't matter with current graphics since the edges are generally transparent, but seem to be a sign that somethings wrong.
  
  * You'll notice that the text in the yellow corner is always backwards. Apparently the image is either mirrored vertically or horizontally, but never not mirrored and never mirrored both ways. We'd get a more natural, irregular effect if each graphic had more than these two positions, especially if 90° rotation was also sometimes thrown into the mix.
  
  * And most importantly to me, the scaling range is too extreme. Sometimes the graphic can be shrunk so small that the star's halo pretty much covers it. This is what got me to investigate, since sometimes big gaps are visible in thin spiral arms. At max zoom the minimum size of a 256x256px gaseous graphic should be about 100%, i.e. 256x256px on screen. If the artist wants blank spots, he can throw some blank files in, but i don't think they look quite right.

I realize there are way more important things that need work, so now that i got this off my chest, i'll go file a FT, and hope somebody gets to it eventually.

As far as I can tell, the gas is supposed to be rendered all the same size, and with arbitrary angle rotations (not just 90 degrees or flips). I think there's a problem with the rotated coordinates calculation math that ends up making it look like a big range of scaling factors and no rotations is being used.

The stretched edges is just what OpenGL does when not tiling or mapping (stretching) a texture that's being drawn on a quad that's bigger than the texture. It can be seen sometimes with system icons as well, if you turn off gas. Can't really do anything about it except avoid using larger quads than textures (not always possible) or putting transparent borders on textures.

Edit: after some fixing, I get: