The Outerra Engine

[gammadust 12]

Equally... but did not do what you just did here - finaly register - i remember when both communities got all heated up about just the remote possibility of cooperation between these techs ;)

I am hearing good news from multiple sources for very distinct reasons > wide smile in me

[minimalaco 30]

Awesome ! downloading now ! :D

[NeMeSiS 11]

Well, as i expected this gives a rather disturbing view of a completely flooded Netherlands. :p

Edited February 27, 2012 by NeMeSiS

[gammadust 12]

Which can only mean the source data is correct.

[whisper 0]

Thanks! Been lurking here occasionally for some time already :)

Joining Gammadust to say hello, and welcome here. Hell of a job you've done so far guys! Hoping the best for your baby

[4 IN 1 0]

The demo don't work for me, crash every time when I try to move my mouse once inside the demo. guess it is my 7970 and leaked RC driver that cost the problem

[NeMeSiS 11]

For all the fuss about this in ArmA3 topics this thread is suprisingly quiet.

[onlyrazor 11]

Maybe people have finally realised that changing game engines mid-development would be a monumentally silly idea and have decided to stop pushing. :rolleyes:

[NeMeSiS 11]

Maybe people have finally realised that changing game engines mid-development would be a monumentally silly idea and have decided to stop pushing. :rolleyes:

You mean that people would start making sense? Dont be ridiculous. :p

[onlyrazor 11]

You mean that people would start making sense? Dont be ridiculous. :p

I'd post a picture with the caption of "Oh, God, he's making sense" but then I googled it and the first result was a picture of Criss Angel. I guess you're right my good sir :clap:

[mr burns 132]

For all the fuss about this in ArmA3 topics this thread is suprisingly quiet.

Well maybe it´s time to change that!

I´d for one be willing to sacrifice any gfx bling we have in ArmA2 for a game that´d combine Outerra scale and model density for something bigger that had clients of all game branches matched into one big ugly mother..destroying game of the year. The future is near! Behold global online play as the fast clicking rts nerds get to command real time manoeuvres to real time WoT, IL2, Steel Beasts, CS or ArmA players!

Me wants one now

[h - 169]

I guess when the thing doesn't even run in full-screen makes it a bit less appealing..

[ballistic09 241]

I guess when the thing doesn't even run in full-screen makes it a bit less appealing..

Works for me...

Took a few tries and a little effort though. Just keep setting it to fullscreen in the settings and restarting it, eventually it'll work.

[noubernou 77]

I'd just prefer that Arma support view distances and object distances greater than 10,000-30,000m with out totally falling apart performance wise or having to make huge sacrifices in terrain fidelity. :p

[celery 8]

Personally, I'd see the best application of Outerra to be in space games where you could take the action planetside (something I really liked in Frontier: Elite II and have missed since), and perhaps some flight simulators by extension. The immense scale and endlessness of available terrain would do little to make an infantry-focused military game better if the terrain and its features are too undetailed for infantry action and storing the data takes gigabytes upon gigabytes of space.

[Tonci87 163]

Personally, I'd see the best application of Outerra to be in space games where you could take the action planetside (something I really liked in Frontier: Elite II and have missed since), and perhaps some flight simulators by extension. The immense scale and endlessness of available terrain would do little to make an infantry-focused military game better if the terrain and its features are too undetailed for infantry action and storing the data takes gigabytes upon gigabytes of space.

This

Imagine a space game where you can do proper bombings from space....

[gammadust 12]

Personally, I'd see the best application of Outerra to be in space games where you could take the action planetside (something I really liked in Frontier: Elite II and have missed since), and perhaps some flight simulators by extension. The immense scale and endlessness of available terrain would do little to make an infantry-focused military game better if the terrain and its features are too undetailed for infantry action and storing the data takes gigabytes upon gigabytes of space.

Are you, by any chance, missing the point about Outerra wherein most detail is proceduraly generated and by that means exacly cutting down the amount of data required to represent such huge landscapes?

Original data used is a sparse 90m grid of data height points, note how the engine cleverly generates both low and high frequency datail to the scene, and also the verisimlitude it sustains. This, imo, is exacly the strength of this engine.

check (original data in the example is even sparser - 150m) follow "Horizontal Displacement Effect" link in the bottom too.

I concede it may be still early with the state of technology, but I would guess this would be very fitting for "infantry-focused" intents, keyword being procedural.

[Dwarden 1125]

Personally, I'd see the best application of Outerra to be in space games where you could take the action planetside (something I really liked in Frontier: Elite II and have missed since), and perhaps some flight simulators by extension. The immense scale and endlessness of available terrain would do little to make an infantry-focused military game better if the terrain and its features are too undetailed for infantry action and storing the data takes gigabytes upon gigabytes of space.

such games and engines exist already ...

e.g. opensource remake of 2nd Frontier (First Encouters , also marked often as Elite 3)

it's name is Pioneer : http://pioneerspacesim.net/

not to mention Outerra so far proved it's capable only of single planetary body rendering ...

there wasn't anything to prove otherwise like star system or star clusters ...

also people not realize that even if You have working engine ready for gaming usage

(which Outerra is still far from) then You need actually make the game for it ;)

---------- Post added at 19:03 ---------- Previous post was at 18:55 ----------

Are you, by any chance, missing the point about Outerra wherein most detail is proceduraly generated and by that means exacly cutting down the amount of data required to represent such huge landscapes?

Original data used is a sparse 90m grid of data height points, note how the engine cleverly generates both low and high frequency datail to the scene, and also the verisimlitude it sustains. This, imo, is exacly the strength of this engine.

check (original data in the example is even sparser - 150m) follow "Horizontal Displacement Effect" link in the bottom too.

I concede it may be still early with the state of technology, but I would guess this would be very fitting for "infantry-focused" intents, keyword being procedural.

you seems to missing the point that size of procedural data storage keeps increasing with level of detail too

you need more detailed seeds covering more complex microdetails

so yes while it's by many folds smaller than any other method except fractals (that seed would be even smaller)

it still become quite huge for believable and usable micro-detail (AI pathing)...

Note: we have procedural infinite terrain in our games for years, same as procedural textures support ...

now comes another problem, dynamic destruction of objects (be it structures or bushes/trees) and then terrain

for each change You would need add seed which would reflect these changes (e.g. crater in ground, broken tree, damaged structure)

so the game-size would grow significantly as You and AI and environmental (e.g. erosion) changes alter the world

[celery 8]

Are you, by any chance, missing the point about Outerra wherein most detail is proceduraly generated and by that means exacly cutting down the amount of data required to represent such huge landscapes?

Original data used is a sparse 90m grid of data height points, note how the engine cleverly generates both low and high frequency datail to the scene, and also the verisimlitude it sustains. This, imo, is exacly the strength of this engine.

check (original data in the example is even sparser - 150m) follow "Horizontal Displacement Effect" link in the bottom too.

I concede it may be still early with the state of technology, but I would guess this would be very fitting for "infantry-focused" intents, keyword being procedural.

Procedurally creating believable and most of all good environments for infantry action becomes feasible only when you can get something like Chernarus or even Everon with the press of a "generate" button. The set of rules for placed objects and their coexistence would have to be astonishingly complex for it to be anything more than a curiosity that sacrifices everything for the fact that it's procedural.

such games and engines exist already ...

e.g. opensource remake of 2nd Frontier (First Encouters , also marked often as Elite 3)

it's name is Pioneer : http://pioneerspacesim.net/

It would have been nice to have a planetary dimension in other games than Elite and its remakes. E.g. Freelancer would have been so much better if it wasn't limited to automatically going through a gate and entering a menu that looks like a spaceport.

[gammadust 12]

No specialist here, no... But it does not have to be a all-or-nothing approach, procedural can coexist with content made from the ground up.

My point is that certain type of content can hugely benefit from procedural generation (in a layered approach - terrain topology, types of terrain, roads, even buildings and tree placement, and i suspect many more if we delve into it - thinking about existing vectorial data of all types)

The compromise is, of course, the generic "feel" this content might impose, i guess though, terrain topology lends itself perfectly.

The point Dwarden raises, about destruction and terrain dynamics, i was not aware...

A city proceduraly generated (yes is not the most organic of the examples)

Recognizing it is not yet "there" i don't think we should dismiss this technology, since basicaly what it does is by automating a lot of content creation relieves artists to concentrate on higher detail stuff.

[Dwarden 1125]

Procedurally creating believable and most of all good environments for infantry action becomes feasible only when you can get something like Chernarus or even Everon with the press of a "generate" button. The set of rules for placed objects and their coexistence would have to be astonishingly complex for it to be anything more than a curiosity that sacrifices everything for the fact that it's procedural.

It would have been nice to have a planetary dimension in other games than Elite and its remakes. E.g. Freelancer would have been so much better if it wasn't limited to automatically going through a gate and entering a menu that looks like a spaceport.

the Pioneer is already more than just remake ...

Freelancer and some other great games like Independence Wars serie universe would really shine if they had procedural engine of Pioneer/Frontier

there is another engine which offers fantastic voyage to fully procedural Universe (let say local supercluster of galaxies size to any planet surface)

while the planetary quality isn't yet on level Outerra or Pioneer in detail it can match early procedural Terragen surfaces easily

each dot is usually star system, which has own system with stellar bodies and you visit each of them ...

all above was taken while rendering realtime on my archaic computer

---

speaking of our engine well we already have 'erm script-dynamic (not sure if procedural)' town generation ...

procedural object placement is possible for vegetation (similar like clutter is done)

you can make specific rules, some of them are part of natural order anyway

e.g. altitude, pressure, oxygen %, humidity, wind strength (too strong cause erosion and enforces what type of trees can alive), water sources (surface, underground), type of soil

by these rules You can then get fully procedural placement of vegetation

similar rules can be applied to rock and other objects placement

procedural cities are already available for Blender and also some games (i think even some simulators use it for cities/villages)

the primary issue is with how find balance between stereotype and natural complexity w/o too big/complex seed/calculations

---------- Post added at 22:32 ---------- Previous post was at 22:19 ----------

@gammadust nobody is dismissing procedural engines, in fact together with voxel engines it's one of ways how most likely game development goes ...

(probably blend of available and these these techs) ...

to reach fully procedural engine on universe scale , you would need combine all the various techs

this mean not just cities, vegetation, terrain, weather/erosion, enviroment (flora/fauna), economy, AI, destruction

but also way how procedural content reacts and alters by players actions and the living world reaction on it

i'm sure we see more use of it in future simulation, strategy and building/contruction games ... (e.g. SimCity like)

---

and i think that Cloud technology would allow to also handle such amount of data and processing demand (interlinked server farms)

[noubernou 77]

Its not that Arma needs a world, its that we need something bigger than most maps we have now. 40x40 is about as big as you are going to get in the A2 engine without dropping terrain quality to something absurd, and at 40x40 forget any sort of decent micro terrain as well.

Playing as infantry is fine on 10x10km maps, but doing full on mechanized warfare where you might be advancing dozens of kilometers in a very short period of time is rather hard to do on most terrains out there. Also being able to really expand out the use of assets like self-propelled artillery and helicopter support would make the game a lot more fun.

Yes, A3 is going to be bigger, but its back to an island again, which feels entirely limiting.

If there was a way to have dynamic terrain like there is in Outerra but limited down to smaller, custom generated height data that would be perfect.

I mean Outerra, compared to Arma, the more terrain you see the higher the FPS you get. I find that rather amazing (and makes me wish that A3 would support similar terrain tessellation).

[gammadust 12]

http://img.ie/images/53833_thumb.jpg http://img.ie/images/13c38_thumb.jpg http://img.ie/images/8b697_thumb.jpg http://img.ie/images/bbc62_thumb.jpg http://img.ie/images/51aa9_thumb.jpg

http://img.ie/images/2011b_thumb.jpg http://img.ie/images/b7bc4_thumb.jpg http://img.ie/images/bb015_thumb.jpg http://img.ie/images/06098_thumb.jpg http://img.ie/images/cc406_thumb.jpg

Btw, which engine was that?

i'm sure we see more use of it in future simulation, strategy and building/contruction games ... (e.g. SimCity like)

No doubt. I remember when MonteCristo (Cities XL original developer) was struggling to put out a number of buildings, i think they reached 4 hundreds, which is considerable and did allow for good diversity.

SimCity5 has been announced btw, I wonder what they'll do...

[Dwarden 1125]

Space Engine: read more http://forums.bistudio.com/showthread.php?132123-Space-Engine

and i made thread also for Pioneer Space Simulator: http://forums.bistudio.com/showthread.php?132121-Pioneer-Space-Simulator

ViewTerra, another full earth simulator: http://forums.bistudio.com/showthread.php?132332-ViewTerra-Earth-in-Your-hands

Edited March 11, 2012 by Dwarden

[hobbesy 10]

http://www.youtube.com/watch?v=WHQdllTxGfI

I'd hate to break the chain of serious discussion, but I thought the most responsible thing to do in the demo would be ramping the truck off the side of the Grand Canyon. Ramping it off to the tune of "The Final Countdown" of course.