improving nebulae and background stars

Development of artwork, requests, suggestions, samples, or if you have artwork to offer. Primarily for the artists.
Message
Author
User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: improving nebulae and background stars

#16 Post by eleazar »

Here's a proof of concept: It's about what i'd like the "gaseous substance" to look like.
The same texture (nebula2.png) is centered over each star and randomly rotated. (of course for the real thing we would have multiple textures).
IMHO the way multiple layers overlap produces a very nicely irregular effect
Attachments
star.jpg
star.jpg (144.31 KiB) Viewed 2210 times
partial example
partial example
gasseous2.jpg (54.03 KiB) Viewed 2247 times

User avatar
pd
Graphics Lead Emeritus
Posts: 1924
Joined: Mon Mar 08, 2004 6:17 pm
Location: 52°16'N 10°31'E

Re: improving nebulae and background stars

#17 Post by pd »

geoff wrote:
pd wrote:
Geoff the Medio wrote:Why do you have background big-nebula-like-texture extending well out into intergalactic space? This distorts the gameplay-shape of the galaxy significantly.
Let's not confuse things. The nebulae have nothing to do with the cloudy things and the pixel stars(let's just call them dust from now on) which shape the galaxy.
I meant the cloudy things, which it appears you've used current nebula textures to create... hence "nebula-like-texture"
You meant the big faint non-gameplay-nebulae, which - as I've said - has nothing to do with shaping the galaxy.
geoff wrote:I mean the stuff outside the cyan circle and enclosed in magenta in the attached.
[attachment]
eleazar wrote:I agree that it looks better and is easier to use if the nebula textures stick a lot closer to where the gameplay stars are.
I wrote:It's just an artistic choice. It's just some overlap, some variance.
I don't want the nebulae to be perfectly circular or circular cropped. I might have gone a bit overboard with this, but it's pretty dark and hard to see anyway and has no gameplay value either. It's quite important when discussing about those things is to have the screen properly calibrated. A fairly easy way to do this - at least in terms of brightness/contrast/gamma is using a picture like this:
Image
Could you please make sure that both rows of 1-8 are recognizable and that all nine tonal steps are evenly changing in brightness? The first 1(first row) and the last 8(second row) have to be equaly easy been seen.
The issue is that you want "dust" that looks like stars. Dust does not look like discrete bright points of light on a black background... that looks like stars, and looks like the background starfield. Dust, or generic galaxy-filling interstellar gas, could be smooth or blotchy like the stuff in the second attachment, extracted from eleazar's mockup. It could also look like the cloudy things.
It could also include some particles, or some kind of matter to avoid the too clean and cloudy looking mockup labeled "just gaseous substance". See how in the next mockup labeled "gaseous slightly toned down + toned down pixel stars", the "dust" nicely breaks up this too soft looking clouds? We can't go from fairly large sized stars to nothing IMO, there should be a step in between, hence the particles...
Then why does it look like stars?
It just some pixels. They are mid grey in this mockup and will probably be even darker in-game. They really could be anything. If I would want them to be stars(gameplay stars) they would be shown like those.
eleazar wrote:But if that texture was at 100% at max zoom in, then it would probably have to be very high res to cover enough space.
Showing them at 100% when zoomed in is my intention. I don't see why they have to be high resolution though. They are gonna placed just like the gaseous, cloudy things, probably even in the same textures. I discarded the idea of using a single image for all pixel stars/clouds to shape the galaxy previously in the other thread. This thread is all about placing tiny sprites by code, as visualized on one of my first mockups here.
eleazar wrote:In regards to your mock-ups, after looking at them for a while, i think you have too many layers. Your "gaseous substance" [...] is visually about the same thing as your large nebulae with these differences:
1) "gaseous substance" follows the shape of the galaxy
2) "gaseous substance" is a uniform dull grey.
Just to be clear:
- It was my intention to do those gaseous, cloudy, galaxy shape following things using sprites from the beginning. Please see my mockup labeled "just gaseous substance:". Right above it I'm talking about 300 instances and I was assuming to use the same technique es shown in the mockup labeled "To visualize this, here is another mockup"
- There is one large faint nebula that colors those gaseous clouds. This is much easier than coloring by code and adds more variety as well.
- A layer based system is a beautiful thing.
Your "gaseous substance"(which is more or less the same idea i was trying to get at with my mock-up of "blue blobs")
I previously wrote:Notice, those squares are supposed to be randomly flipped and rotated sprites include about 10-30 tiny stars(merly dots) each. We might also introduce some faint gaseous substance, especially for the sprites used in the center of the galaxy.
So yeah, it's the same and it always was.
the negatives:
1) excessively over-zoomed textures, and
2) galaxy boundary trespass.
1)The large faint background nebula is without any definition. Bluring introduced by zooming is not really visible because everything is made of large single colored areas or falloffs anyway. See this old screenshot, when nebulae have been bigger than they are now. This is about how it should look when zoomed in. In addition there will be some more detailed cloud shapes, introduced by the sprites.
2)
I wrote:It's just an artistic choice. It's just some overlap, some variance. I don't want the nebulae to be perfectly circular or circular cropped.
Every further discussion about pixel stars/dust/particles/whatever is useless, because we are running in circles and I won't invest any further time in it until we have some code to play with.
I wrote:Ultimately, I just would like to have some textures placed like in one of my previous mockups shown. What we put into those textures(clouds/pixels) can be fiddled with later.

User avatar
pd
Graphics Lead Emeritus
Posts: 1924
Joined: Mon Mar 08, 2004 6:17 pm
Location: 52°16'N 10°31'E

Re: improving nebulae and background stars

#18 Post by pd »

Looks great eleazar. The halos/spikes could need some more work to avoid the regularity. And the stars should either get smaller, or the distances larger IMO. But that's another topic.

In my previous mockups I tried to avoid those holes between stars. it might look good with just a small part of an galaxy arm covered in sprites, but when filling the entire galaxy those holes look weird. We could avoid this by either using larger sprites or by filing the triangles in between as well.

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: improving nebulae and background stars

#19 Post by eleazar »

Well, i think we've accomplished a good deal today. :)

pd wrote:You meant the big faint non-gameplay-nebulae, which - as I've said - has nothing to do with shaping the galaxy.
It's obvious that in your mock-up they don't have anything to do with the shape of the galaxy. But the point i've made, and i think Geoff is making is: whatever nebulous-type-textures we use should follow (more or less) the shape of the galaxy.
me wrote:. . . extending nebulas far beyond where the player can go... these appear much more definitely like a "place" where the player might want to go. That's what what we need to avoid... as long as all the interesting-looking stuff is part of our playable galaxy, that limitation will be much less likely to be noticed.
... additionally it makes it easier to understand where you are in the galaxy if the nebula-texture-stuff more or corresponds to the part of the map where the game takes place.
pd wrote:It's just an artistic choice. It's just some overlap, some variance.
I don't want the nebulae to be perfectly circular or circular cropped. I might have gone a bit overboard with this, but it's pretty dark and hard to see anyway and has no gameplay value either.[/quote]
I really don't think geoff want's a circle crop. I don't find that artistic choice very compelling, especially when it has a somewhat negative effect on ease of use. IMHO there's not a way (with your giant nebulae) to make them conform even as well as they do in the mock-up to the galaxy shape. In many instances they would probably end up placed in a much less galaxy-conforming location.
pd wrote:Could you please make sure that both rows of 1-8 are recognizable and that all nine tonal steps are evenly changing in brightness? The first 1(first row) and the last 8(second row) have to be equaly easy been seen.
Yes, all steps are quite clear and close to being equal steps.
pd wrote:
eleazar wrote:In regards to your mock-ups, after looking at them for a while, i think you have too many layers. Your "gaseous substance" [...] is visually about the same thing as your large nebulae with these differences:
1) "gaseous substance" follows the shape of the galaxy
2) "gaseous substance" is a uniform dull grey.
Just to be clear:
- It was my intention to do those gaseous, cloudy, galaxy shape following things using sprites from the beginning. Please see my mockup labeled "just gaseous substance:". Right above it I'm talking about 300 instances and I was assuming to use the same technique es shown in the mockup labeled "To visualize this, here is another mockup"
- There is one large faint nebula that colors those gaseous clouds. This is much easier than coloring by code and adds more variety as well.
- A layer based system is a beautiful thing.
I see and understand your mockup, and what you intend all the different layers to be. And i still think it's too much. In your mock-up giant nebulae almost completely cover the "gaseous substance". Why do we need them both? They both provide basically the same visual effect, except the hundreds of sprites can be much more easily controlled.
pd wrote:
me wrote:the negatives:
1) excessively over-zoomed textures, and
2) galaxy boundary trespass.
1)The large faint background nebula is without any definition. Bluring introduced by zooming is not really visible because everything is made of large single colored areas or falloffs anyway. See this old screenshot, when nebulae have been bigger than they are now. This is about how it should look when zoomed in. In addition there will be some more detailed cloud shapes, introduced by the sprites.
OK, i grant you #1.
But now that we (probably) have a ready means to make pretty good looking nebula-like-textures, i see no compelling reason to put up with "galaxy boundary trespass."

pd wrote:In my previous mockups I tried to avoid those holes between stars. it might look good with just a small part of an galaxy arm covered in sprites, but when filling the entire galaxy those holes look weird. We could avoid this by either using larger sprites or by filing the triangles in between as well.
It's a lot easier to avoid things like "holes" when your doing things by hand. Well, yes technically i made that mock-up by hand, but i just dropped the same graphic over and over, to replicate the process. Holes will happen, especially in irregular galaxies. But i do agree that the sprites should probably be larger than in my example. I just didn't want to change all those graphics.

User avatar
pd
Graphics Lead Emeritus
Posts: 1924
Joined: Mon Mar 08, 2004 6:17 pm
Location: 52°16'N 10°31'E

Re: improving nebulae and background stars

#20 Post by pd »

I think we do agree about the major points.
We both want to shape the galaxy using hundreds of sprites, located at stars or in between them.
Whether those include pixel stars is not important for now. That's the one thing.

Additionally, unrelated to shaping the galaxy, I want to unify the appearance of the galaxy in terms of nebulae, which are just randomly colored blobs now. This is done by using a fairly consistent color scheme and having a huge FAINT nebulae connecting them. The main purpose of this is, to still have small distinct areas, so that certain stars could be effected in a certain way. Additionally it comes in handy with coloring the sprites.

Obviously there has to be some logic in placing(and scaling) the large nebula, so that is occupies a major part of the galaxy.
IMHO there's not a way (with your giant nebulae) to make them conform even as well as they do in the mock-up to the galaxy shape. In many instances they would probably end up placed in a much less galaxy-conforming location.
That's just a nebula we are already using - placed not quite in the center of the map. There is nothing masked out and retrospective I should have probably been more careful with placing it, because we could have avoided a lot of this discussion.

You want to abstain from the nebula and color the sprites instead by code. That's the point where we differ. I believe we have more control about how to color the sprites using a single large picture, instead of tweaking certain numeric values. We could for example easily have all kind of different patterns with a picture overlay. Wanna play in a smiley galaxy? No problem, create a smiley nebula. I don't say we should do this, I just say we could. Doing something like this by code is quite hard I imagine(not the smiley alone, but the whole palette of possibilities).

Well, i think we've accomplished a good deal today.
Hell yes, we have :)

User avatar
Flatline
Space Squid
Posts: 64
Joined: Tue Jan 29, 2008 2:06 pm
Location: Oxford (UK)

Re: improving nebulae and background stars

#21 Post by Flatline »

eleazar wrote:Here's a proof of concept: It's about what i'd like the "gaseous substance" to look like.
The same texture (nebula2.png) is centered over each star and randomly rotated. (of course for the real thing we would have multiple textures).
IMHO the way multiple layers overlap produces a very nicely irregular effect
I really like the global effect (besides intensity of the color, "background nebula" should be a little less evident in order to make "game nebula" easy to recognize).

About multiple textures: it is ok, but they should all have almost the same colour (honestly in some previous posts I've seen a merge of red and blue "gaseous substance" which I didn't really like), but whe could choose randomly which one they have at game start (and the nebula png could just be a b&w alpha which we colorize by software).
The only difference between a suicide and a martyrdom is press coverage.
- Chuck Palahniuk (Survivor)

Ubuntu 7.10/Windows Vista/Windows XP (VMWare)/Windows NT 4 (VMWare) on Intel Centrino Duo [email protected] with NVidia GeForce GO 7600

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: improving nebulae and background stars

#22 Post by eleazar »

pd wrote:Obviously there has to be some logic in placing(and scaling) the large nebula, so that is occupies a major part of the galaxy.
• Yeah, as Silent mentioned earlier the scaling factor of a nebula should be linked to the size of the galaxy. This would of course apply to your "huge faint" nebulae or small ones.

• Additionally to keep them within the galaxy, a simple (but not foolproof) method would be to place nebulae centered only at the coordinates of a star-system. This would provide better results than we now get, but maybe somebody could come up with something even smarter.

pd wrote:
eleazar wrote:]IMHO there's not a way (with your giant nebulae) to make them conform even as well as they do in the mock-up to the galaxy shape. In many instances they would probably end up placed in a much less galaxy-conforming location.
That's just a nebula we are already using - placed not quite in the center of the map. There is nothing masked out and retrospective I should have probably been more careful with placing it, because we could have avoided a lot of this discussion.
No, you only would have postponed the discussion. If you had carefully hand-placed the nebula to conform to the galaxy shape, and/or masked out unwanted bits, it would have simply hidden something i consider a problem until i thought it through or saw it actually working.

pd wrote:You want to abstain from the nebula and color the sprites instead by code. That's the point where we differ. I believe we have more control about how to color the sprites using a single large picture, instead of tweaking certain numeric values.
No, you misunderstand. I don't want FO to colorize the graphics— that's control i too would rather exercise myself. I suggested that if we wanted multi-color galaxies (not something i'm entirely sold on, but it seems to be important to you) we could code the universe generator to usually choose sprites of a similar/same color to adjacent sprites... thus if color changes would generally cover large portions of the galaxy rather than be patchwork.

Here's the process i see:
  • 1) implement gaseous substance with a single set of similarly colored graphics. (i'll probably make some working graphics today)
    2) implement multiple sets of graphics, but only use one per galaxy
    3) ? consider ways to use multiple colors without it looking bad
If we only get to #2 we've still improved things quite a bit.

pd wrote:We could for example easily have all kind of different patterns with a picture overlay. Wanna play in a smiley galaxy? No problem, create a smiley nebula. I don't say we should do this, I just say we could. Doing something like this by code is quite hard I imagine(not the smiley alone, but the whole palette of possibilities).
I realize you are joking about the smily face, but i don't understand what these possibilities are. I don't see why we would want to add a giant graphic to the galaxy which had nothing to do with the shape... or what do you mean?

Flatline wrote:..."background nebula" should be a little less evident in order to make "game nebula" easy to recognize).
It's important to remember that there are no "game nebulae". Some people would like to add that concept to the game, but i'm not going to spend much sweat trying to support an indefinite feature which hasn't undergone any serious discussion.

Flatline wrote:About multiple textures: it is ok, but they should all have almost the same colour (honestly in some previous posts I've seen a merge of red and blue "gaseous substance" which I didn't really like), but whe could choose randomly which one they have at game start...
Generally i'd like to see things harmoniously colored... but we can't go too monochrome. The only gameplay function the nebulae currently serve is as "landmarks" to help distinguish different parts of our unusually-large-for-a-game galaxies. If all the non-star graphics are pretty much the same color, the "landmark" function will be impaired. I'm assuming that the "small" nebulae will serve most of the "landmark" function.

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: improving nebulae and background stars

#23 Post by eleazar »

I've added some graphics to test with in art/galaxy_decorations/...

I recommend starting at 30% opacity, and scaling so that at max zoom the sprite width/hight is at 500%.

I'm curious as to how much color variation we can include in the graphics. From my mock-ups it looks like the variation i've included in the files is barely noticeable.



EDIT:
I think i figured a more reliable way to identify the "middle" of an arbitrarily shaped galaxy, for purposes of making the "middle" brighter (as in a normal galaxy) than the "edges".
Assuming the game can identify stars on the "outside edge" of the starlane system:
1) find the shortest number of jumps from every star to the nearest "outside edge" star. Remember that number (SNJ#) for each star.
2) The highest SNJ# found is used to calibrate the series.
3) each star is assigned it's "gaseous substance" image transparency, or the the image from a continuum is chosen based on where the star's SNJ# is between 0 and the max found in #2.

User avatar
pd
Graphics Lead Emeritus
Posts: 1924
Joined: Mon Mar 08, 2004 6:17 pm
Location: 52°16'N 10°31'E

Re: improving nebulae and background stars

#24 Post by pd »

eleazar wrote:we could code the universe generator to usually choose sprites of a similar/same color to adjacent sprites... thus if color changes would generally cover large portions of the galaxy rather than be patchwork.
That's definitely possible. In this case you would have multiple sets of those gaseous sprites, right? And you will need some code to choose from those appropriately. That's IMO loss of some control.

While my proposed system uses just one greyscale set of gaseous sprites and one huge faint nebula kind of thing, placed in the middle(yes, in the middle) of the galaxy and scaled to fit the galaxy size to color the sprites. The nebula is so faint, that it is barely visible on black background, but adds up nicely when occupying underlying sprites. This nebula is roughly circular shaped - this way it works with adequate scaling for all galaxy shapes. In a rectangular galaxy just scale it bigger or leave the edges uncolored or take a rectangular shaped nebula. I believe this approach is easier to implement, less work for the graphics people and it provides more control at the same time.
Here's the process i see:
  • 1) implement gaseous substance with a single set of similarly colored graphics. (i'll probably make some working graphics today)
I quote this just to emphasize it. Should we create a sourceforge feature request or is this done internally?

I realize you are joking about the smily face, but i don't understand what these possibilities are. I don't see why we would want to add a giant graphic to the galaxy which had nothing to do with the shape... or what do you mean?
I don't know what to do with in terms of game play, but I like the idea of having full control over the color of a galaxy. I can easily determine which part has what color, just by creating an image. Colors can change from the inside to the outside, from top to bottom, in a checkerboard pattern or everything you can think of. It's really just about not limiting ourselves.
It's important to remember that there are no "game nebulae". Some people would like to add that concept to the game, but i'm not going to spend much sweat trying to support an indefinite feature which hasn't undergone any serious discussion.
That's right and perfectly reasonable, but if I've learned one thing from this project, it is to work with solutions that are easily to change later. We have to think about what could happen later most of the time.
Flatline wrote:[...)in some previous posts I've seen a merge of red and blue "gaseous substance" which I didn't really like).
We really can't consider personal color preferences. From an artistic point of view red and blue work well together, if one is kept superior and the other one is used as an accent. This has to do with the concept of unity and variance, which aplies to many things in art(color, proportion, shape language, line weight,...)
Generally i'd like to see things harmoniously colored... but we can't go too monochrome.
Exactly.

User avatar
pd
Graphics Lead Emeritus
Posts: 1924
Joined: Mon Mar 08, 2004 6:17 pm
Location: 52°16'N 10°31'E

Re: improving nebulae and background stars

#25 Post by pd »

eleazar wrote:I've added some graphics to test with in art/galaxy_decorations/...
Thanks, this should get the coders started. It's perhaps a bit to detailed, which might make repetition obvious. Fortunately, creating or changing those is a no-brainer and I think, we can achieve some interesting effects with just using different kinds of cloudy structures
I think i figured a more reliable way to identify the "middle" of an arbitrarily shaped galaxy, for purposes of making the "middle" brighter (as in a normal galaxy) than the "edges".
I'm sure the exact center of a galaxy is either clear from the beginning or can easily determined at the beginning, even before the creation of stars and star lanes start. I mean, there have to be some coordinates and dimensions the algorithms start working with, right?

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 13587
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: improving nebulae and background stars

#26 Post by Geoff the Medio »

See attached.

The yellow tiny star texture seems to be lacking transparency.

Note the FPS... Apparently 500 textures per frame at that size does add up.

Also, I had to reduce the alpha of the gassy texture to 50 / 255 to get this look. Higher alphas were more opaque.
Attachments
galaxy with gassy substance and tiny stars
galaxy with gassy substance and tiny stars
fo_gassy_w_tiny_stars.png (685.09 KiB) Viewed 2126 times

tzlaine
Programming Lead Emeritus
Posts: 1092
Joined: Thu Jun 26, 2003 1:33 pm

Re: improving nebulae and background stars

#27 Post by tzlaine »

FWIW, we can probably beat this FPS figure by quite a lot by rewriting the starmap rendering code. We currenty don't batch anything, but we could batch everything. This would probably have been needed for large maps even without these new decorations.

User avatar
eleazar
Design & Graphics Lead Emeritus
Posts: 3858
Joined: Sat Sep 23, 2006 7:09 pm
Location: USA — midwest

Re: improving nebulae and background stars

#28 Post by eleazar »

Well, if we can take care of the FPS issue, i think this demonstrates we can get what we want with this method.
Geoff the Medio wrote:The yellow tiny star texture seems to be lacking transparency.
that's a remarkable uglification from only 4 pixels per star.
fixed.
Geoff the Medio wrote:Note the FPS... Apparently 500 textures per frame at that size does add up.
:( I knocked them back to 128x128. And that's 5 fps without any rotation of the texture, if i'm not mistaken.
If that's not enough, you might want to try using only one of the gaseous textures.
Geoff the Medio wrote:Also, I had to reduce the alpha of the gassy texture to 50 / 255 to get this look. Higher alphas were more opaque.

User avatar
Geoff the Medio
Programming, Design, Admin
Posts: 13587
Joined: Wed Oct 08, 2003 1:33 am
Location: Munich

Re: improving nebulae and background stars

#29 Post by Geoff the Medio »

Cluster galaxy, with fixed yellow tinies.
Attachments
cluster galaxy with gassy substance test
cluster galaxy with gassy substance test
fo_gassy_w_tiny_stars_cluster.png (730.78 KiB) Viewed 2111 times

User avatar
Flatline
Space Squid
Posts: 64
Joined: Tue Jan 29, 2008 2:06 pm
Location: Oxford (UK)

Re: improving nebulae and background stars

#30 Post by Flatline »

pd wrote:
Flatline wrote:[...)in some previous posts I've seen a merge of red and blue "gaseous substance" which I didn't really like).
We really can't consider personal color preferences. From an artistic point of view red and blue work well together, if one is kept superior and the other one is used as an accent. This has to do with the concept of unity and variance, which aplies to many things in art(color, proportion, shape language, line weight,...)
Sorry, I didn't mean that I didn't like the two colours used... I meant that I like more a "monochromatic" galaxy :)
The only difference between a suicide and a martyrdom is press coverage.
- Chuck Palahniuk (Survivor)

Ubuntu 7.10/Windows Vista/Windows XP (VMWare)/Windows NT 4 (VMWare) on Intel Centrino Duo [email protected] with NVidia GeForce GO 7600

Post Reply