[POLISHING] Coral Reefs

[Koniption]

Originally posted by morcroft
Right - found it.

CLAIM: City Planning and AI

From what KP says, pathgrids underwater are A Bad Thing, but that without them fish will tend to try and seek out pathgrids on land. The best plan is to have the fish seek out rocks, but that will mean scripting the fish, I suppose.

Hey guys, how goes things? Ya know, Valenwood might be able to be referred to as a sort of "Resurrection Plant" mod - it dies, it lives, it dies, it lives...etc. hehehe

About not using pathgrids underwater because they impact performance - this is something I heard about or read about on some modding-related forum. I have not tested using pathgrids myself. So if you trust my memory recall (which sometimes, I don't trust myself), then don't use pathgrids.

I could very well be recalling my info wrong (insert "Disclaimer" here), but I think it kind of makes sense that adding too many extra "nodes" under water for pathing, will start to clutter the waterways. Like said in above post linky, just have fish seek out the most commonly used underwater object, whether coral reef, coral, rock, or other.

If you want good, comprehensive scripting ideas for the fish, I recommend contacting "Setsuna515" over at The Wormhole:

http://thewormhole.nfshost.com/forum/

She's absorbed the knowledge of various ways to script fish to move that various mods have already tried (including using some ideas of mine, from when I was trying to get fish to move underwater on their own - I had come up with some novel solutions, that did not involve scripting, hint.). She's processed that info, and then spit out a what surely is the most performance-friendly and effective way to get fish to move and act realistic underwater (again, a "Disclaimer", heh).

She works on Craftybits mod team. You'll have to register to PM her over at The Wormhole.

Koniption

[drmccoy]

Hi Kon,

I'm positive I'm able to script the fish by myself, but the more people contribute to this mod, the better it will hopefully become. Plus I won't have to play along with the fishes for too long. The reason the pathgrids shouldn't be used underwater is actually not a performance issue but found in the engine. Gamebryo isn't really laied out for massivly populated watermasses. Small waypoint-networks for little ponds or similar sizes aren't a problem, but the larger they become, the more problems they will cause. You see, even though the pathgrids are still placed on the same solid ground as they would be on a landmass, the watermass surround them is calculated aswell (you might think of it as nether or heavy fog). This leads to the phenomena of Gamebryo Engine wanting to create a 3 dimensional network of pathgrids, while in the editor, you can really only specify 2 dimensions. Of course, you can alter it's z-position, but the waypoint itself will be compiled right above the ground. Not so under water. .... aaanyways, I'm losing myself. Point is, as long as you're underwater, Gamebryo tries to interpolate a 3 dimensional pathgrid network, which leads to heavy perfomance issues. Or, lets say, it CAN lead.

Btw: The Fish already active in the Mod seem to move? I'll check our their script tomorrow, I'm really tired by now

[morcroft]

Just taken my first ever look at these islands in the CS - good job so far everyone (past and present) :goodjob:

If it's OK with you guys, I thought I might have a go at running a big arc round the south of the middle island, from about (-3, 1) to the existing reef at (0, 1), and maybe straying out as far as (-1, 0). Not continuous, of course, but noticeably linked up. That leaves me well clear of Deeza's Small island reef and he can continue across the north of the island while I plod along in my usual slow fashion....

I assume it's mostly just variations on the theme of what's there at the moment.

The 0, 1 reef doesn't have any vertex shading under it - I'm guessing it should. I'll include that in my claim.

[EDIT]
Oh, and I guess we should make sure to remember to include at least some gaps for shipping lanes!

[Deeza]

Sounds like a great plan to me. I'll continue to work on the small island and you can carry on with your part. :check:

[morcroft]

I think the reef arc idea is working.

I've made the following style decisions, based on the fairly fundamental idea that coral grows underwater (!)

Valenwood reefs are made (exclusively) of the four Coral Rock meshes, which seem to work fine scaled anywhere between about 0.8 and 1.5.

Coral does not stick out of the water - ever. It's not like a plant and it can't possibly grow that way.

Coral grows on solid foundations, so somewhere under the reef is rock. Or shipwrecks. Whatever. Anywhere the reef sticks out of the water it should be bare rock.

All reef-associated rock is the ValenwoodBeach rock set.

More as a style idea, I've decided coral doesn't like to grow on the shoreward side of rocks in the shallows - maybe tides and surf wash it off. Anyway that gives me the excuse to have bare rock near the shore merging into reef on the seaward side. That ties the sea meshes visually to those on land. This doesn't apply more than 10 metres or so from the shore, from there out every rock would be encased in coral so the beach rock isn't visible (ie not placed).

[drmccoy]

If you take a closer look at some isles in the south sea, you will recognize a certain terrain pattern aswell. From above, the water around the isles is a very light blue. That's caused by the depth of the water, which is at the most 3-5 meters. After a while, there's a sharp hilldown and the depth increases rapidly. Look at some pictures, you'll see what I mean. I don't think we'll be able to generate that kind of beauty with the engine, since gamebryo displays every water surface in the same color, no matter about the depth. In general, the light-blue water area is filled with reefs, not the deeper area. Maybe you could use some of those pictures as a "hint" to how far the reefs should be constructed. Also, maybe we could edit the terrain later on a little bit so it falls off after a certain distance. Might be neat.

Cheers :goodjob:

[morcroft]

Another thing - are we keeping the resource file frozen now, or can I create new meshes? Don't plan much, but I'd like to add some barnacles, and I don't think there's either a stock mesh or a suitable valenwood specific one.

[Deeza]

If you are interested in making some, it would be very much appreciated, both here and also in Black Marsh too.

[morcroft]

Yeah - no probs. I've just checked over the nifs from TES3 - there are six different barnacle meshes having varying numbers of a few basic shapes and a couple of small textures. Shouldn't be much trouble to make a set at least as good as that!
You may go mental placing them all though....

[morcroft]

I've now done enough of the reef arc to get some idea of what it's going to look like - so I thought I'd put up some screenies to see what the team thinks, and get any input before I do too much more.





I wanted to prove I could actually mod in the cs, using existing resources and resist the urge to open Blender and model something. Nope. Can't be done. So far I've made some small coral rock "patches" to cover up gaps and the more unrealistic UV areas on the existing coral rock meshes; 6 sets of "barnacles" (more like limpets, really - but TES3 called its version barnacles); a large branched orange coral and a purple starfish.



The player is not the first to be shipwrecked in these treacherous waters...



And this is the overview of the area I've been working on.


I still need to add more coral, including adding a couple more shapes of that orange thing - if people like it. I also need to vary the sand texture so that it doesn't "ripple" around the rocks.

I'll add some more fish - although there are more than really showed up when I did the screenes. They do move quite realistically, by the way doc: no real worries on that score.

So - is this on the right track?