Outerra forum
Outerra Engine => Ideas & Suggestions & Questions => Topic started by: ZeosPantera on December 28, 2010, 01:42:19 am
-
We will see if this topic sticks. I have shown around the 2010 in-review video to several (millions) of people and a few of them have asked water questions. This will help fill the Faq.
Will It...
Flow (If the ground is deformable would water seek the lowest level? Will rivers work that way?)
Displace (If I drive the tatra into a small pond will it overflow because of the volume displaced?)
Have Tides (We have seen the shore video and I presume a general raising and lowering of the tides)
Be removable (can I pump it? pick it up in a bucket?)
-
Need power? Install a Dam and watch hilarity ensue... :)
g.
-
Most of these questions require a more complicated simulation of water, but there could be some simplified approaches.
Flow - rivers will be defined as vectors and water will flow in the direction of these vectors. Water seeking the lowest level - that would be considerably harder to do, especially on a global level.
Displace - could be simulated, but I think this static effect will be insignificant in comparison to dynamic effects that need to be simulated there - splash of water and directional waves.
Tides - can be done if there exists a model describing the changes. Or an approximation using positions of Moon and Sun.
Removable - can be simulated. Basically, lakes will be defined as a 2D polygon and a height of water surface above sea. The height can be adjustable, so if there's an approximation of how the volume changes with height, one can make a fecal truck or whatever they need it for :)
-
Hi there.
I have qusetion closely related with water ;)
What about underwater world? Submarines, scuba diving, exploring deepest deeps of our and other planets?
Did you think about adding this possibility to your engine? How much is it possible to see that feature? (Original engine in-build or allowed by mods etc.)
Thanks forward for answer.
-
Undersea terrain data are included as well, it's possible to dive down already. It will need a proper underwater lighting and physics for submarines. Something will be surely included to demonstrate this aspect.
-
you'll be doing a demo submarine do? I'd really wanted me to be in the demo was a submarine.
-
Nice...
-
So i've seen alot of videos and even Unreal has some dynamic water solutions, alot of these are pretty much particle simulations, like air and gasses, except with gravity so they fall and collide with objects. Theres a game on the Xbox360 Arcade i believe called "Hydrophobia," an Independent title with an engine built solely around a dynamic water simulation about a tremendous city sized ship, naturally, sinking.
Anyways, i was thinking, back in Ued2 and the version of Ued3 used for Unreal Tournament 3, their were displacement surfaces for the likes of water. Simply a sheet divide into tris that would ripple and sway as people ran through them, and it really used very little resources whatsoever. Perhaps if you took a level of detail approach, sort of combined the particle simulations with this displacement..
.. I'm thinking of the old Soft Body physics demo for UDK featuring Marcus rolling around a big green blob of goop. So lets say theres little blobs of goop that can merge with other blobs, and each blob is a based off of a single 'particle'. perhaps blobs can change size depending on how much 'water' is in an area, and when splashed or pumped out, they can subdivide into more particles?
Does this make sense? Kinda reminds me of the slimes in minecraft, lol.
-
(kind of off topic) maybe have a waterwheel to produce power from the river
-
This looks really good!
I am especially impressed with some of the finer details such as what looks like sand being kicked up by the breaking waves on the beach (around 1:50 in the video) :o. This could easily be one of the best real time rendition of water, good job.
I registered just so I could comment on this! as I have a couple of comments.
First of all the waves seem a bit too random... From reading your blog post it looks like this can be fixed with a properly blowing wind.
Related to this is the concept of "Fetch" and I didn't see any reference to it in your blog post (see: http://en.wikipedia.org/wiki/Fetch_(geography) It is a key parameter in modelling wave height that would be nice to see taken into account.
Lastly, and this is my biggest issue with computer water as it seems it is alway pretty much wrong but water is usually wayyyyy too reflective and specular. It seems like you have only implemented sun reflections at this point (and not terrain/objects) so this might be a good time as the sun reflections seem to bright and well defined as they are.
Real world water can be like a mirror but only on very calm days and/or small bodies of water. Most of the time, even light winds cause ripples (at the cm/mm level) on the surface that scatter/diffuse reflections and these seem to be almost always ignored.
These may be small but they make the water look radically different from what the physics of reflection/refraction would predict when modelling a water surface with waves.
Anyways, I believe this should be solvable with today's high power shader programming etc but I have yet to see water rendered in a convincing way... It looks like you have the kind of attention to detail that could make it happen...
Here's a video showing a lot of sea/waves in moderate winds and if you pay attention you will notice you really don't see much reflections of anything (except at the beginning where the water is very smooth). And when you do it is more of a blob of color http://www.youtube.com/watch?v=Lp0Z1SzPJw4&feature=player_detailpage. It's a good example because the orange color offers good contrast with the water surface.
Specifically as well, the islands do NOT reflect on the water surface (which pretty much every rendered sea scape will show...). Even on a calm day you only get some small amount of the color showing like this: http://www.dailyventure.com/photo.php?name=corfu_shore_02&size=1280x1024
Anyways, I hope I don't come across as harsh, I definitely like what I am seeing, I am just trying to offer some constructive criticism
Sorry for the long post and thanks for listening! Keep up the good work!
-
Hi Airwick. Excellent video of water/waves/reflections. I do approve of your suggestions. I always find most game publishers exaggerate the "wow" aspects of everything. Water is mirror shiny, fire is taller, smoke is denser, lens flare f___king exists... etc. These are things I hope to force Cameni to keep out of Outerra with constant yelling on the forums. For the simulators sake. For our sake. For the greater good.
-
Welcome here, Airwick.
There's no concept of "Fetch" as that is a global phenomenon, whereas the water algo now only looks at the local data. For the fetch we'll need the dataset to contain at least the information about the distance to the nearest land mass. It will also become usable only after there's a global wind simulation or data.
Here's a video showing a lot of sea/waves in moderate winds and if you pay attention you will notice you really don't see much reflections of anything (except at the beginning where the water is very smooth). And when you do it is more of a blob of color http://www.youtube.com/watch?v=Lp0Z1SzPJw4&feature=player_detailpage.
Um, on the contrary - the video pretty well shows how water reflects the sky, it's everywhere I look. I can easily see the reflection of sun behind thin clouds there too. Perhaps you are referring to scarce reflections of objects - but that's logical (see below). When comparing the simulation with reality, we must take care we are comparing the same settings. But it is true that the reflections aren't entirely right now even for a clear sky, I'm aware of some things I didn't implement there yet :)
Reflection of terrain and objects - you know, there's a reason for why it doesn't show up much, except when it does.
When we are looking at the waves, we are mostly seeing their front faces. This fact has two consequences - the reflectivity is lower because the overall incidence angle is lower compared to a quiet water, and the reflected ray hits mostly the sky and not the terrain. Of course, this changes if it's a bay and the wind is shielded etc. The problem is that the algorithm must work in all cases, and it's not trivial if it should be also reasonably fast. In the end it's about searching for approximations and endless tweaking. I'm sure we will return to this in the future, but first we have to release something to make it into the future at all :)
Thanks for the post.
-
I spent a little time searching for shots of In-Game water. Some look good, some obviously bad.
http://www.thebuzzmedia.com/wp-content/uploads/2007/11/crysis_screenshot.jpg
http://i94.photobucket.com/albums/l106/Dragonfirexz95/crysis2007-12-1314-00-22-89.jpg
http://i141.photobucket.com/albums/r50/Tieno360/08.jpg
http://i172.photobucket.com/albums/w39/Pingusten/CAPTAIN.jpg
http://www6.incrysis.com/screenshots/31011106cb4.jpg
http://www.pcgames.de/screenshots/original/2008/11/crysis_081108141911.jpg
http://oi44.tinypic.com/2ymaq13.jpg
http://www.pcgameshardware.com/screenshots/original/2009/06/Arma2-12.jpg
-
I like water reflections water in X-plane. They are not excellent, but still mountain reflections are awesome. Start watching from 4:20 minute --> http://www.youtube.com/watch?v=t9tM6k8dusk
And of course REX tropical waves for FSX! I love this kind of color and transparent water! :)
http://www.youtube.com/watch?v=PiPWNVqFUg4
-
I like water reflections water in X-plane. They are not excellent, but still mountain reflections are awesome. Start watching from 4:20 minute
This is actually a really good example of what bugs me: the nearby reflections are decent although too shiny already but the fact that the mountains reflect perfectly is just plain wrong! (unless there was absolutely no wind but then the nearby reflections should be the same).
I read about this on the X-plane forums and the issue there is that the rendering algorithm basically computes the reflections after dealing with the waves and the waves at a distance approximated to have zero height, hence the perfect mirror (wrong) effect.
As for the FSX one, it looks pretty good in that video but I would like to see it at different angles/distances etc. The corral representation is really good but the water surface in FSX is generally too shiny as well unless the reflection details are turned DOWN a few notches...
FSX is in genera very much "in your face" with everything being exaggerated (over-saturated colors/contrast) excessive reflections etc... Don't get me wrong, I do think it looks good and I understand why they would go that way to appeal to the masses (and at least it is done in a coherent way) but it is not what I would call realistic.
Um, on the contrary - the video pretty well shows how water reflects the sky, it's everywhere I look.
I guess I should have been more specific: the sky and "light/colors" do reflect on the water surface but what I meant is that shapes and objects basically don't for the reason that you explain (what we see is mostly the angled face of the waves and not a flat water surface).
X-plane is pretty close to getting it right in the foreground except for the fact that it is missing another level of waves scale. i.e. the faces of ~10 cm waves are smooth whereas in reality they have little wavelets on them that further blur the reflections but the far reflections are completely wrong!
The bottom line is that I was trying to say that I was pleased to NOT see the terrain reflect on the water surface in the videos and screenshots and I hope it stays that way (except for the occasional conditions in which it should)
-
yes some reflections can be over bearing regarless of the weather (original fsx)
-
Tides - can be done if there exists a model describing the changes. Or an approximation using positions of Moon and Sun
I was going to suggest that; good thing I read the thread ;) ...
Tides can have a really big impact on the appearance of landscape etc and would definitely be a nice thing to see in something like Outerra (see the first and last pictures here: http://thecontaminated.com/mount-saint-michael-french-island-and-castle/)
An approximation with the sun/moon position would work well for "open" bodies of water, however it breaks down when you have inlets and closed bays...
I wonder if you could us xtide: http://www.flaterco.com/xtide/xtide.html. It would provide much more accurate data!