First of all, I have been following the development some time and am totally looking forward to your simulation. Finally, nap of the earth chopper flying will be possible with head controlled FLIR and night vision gadgets, yay! I like your open development approach and how you tackle design problems, realistic but approachable and focusing on the mood of the game. And there is nothing like a dynamic campaign when it comes to mood in such a kind of game. Things can always be emphasized with scripts if needed.
Now to something more useful, hopefully.
This probably comes way too late considering the state of development and I don't know if it would be an improvement quality-wise or more practical, also no idea what it would cost or if some co-operation would be possible, still the scenery and view distance in
looks interesting. But I guess it would be difficult to integrate with the existing rendering infrastructure. Rendering only the terrain with that engine might not be easy either. Also if you can fly through trees (which I don't know!) it's no use either.
The Outerra engine has actually been mentioned quite a lot during the development of Combat-Helo, and as far as I'm aware has at least been considered for future game releases from the developers.
Also of note is that link to an open source flight dynamics model - www.jsbsim.org
Global sized rendering is on the cards for possible sequel. Already been looking at technologies for that since I have a hankering to re-visit historic Desert Storm Apache engagements and the Fulda Gap scenario.
Very early on I got in touch with the gentleman, very nice man, behind Outerra. It wasn't available to us and for the project it probably wouldn't have been a good choice for Combat-Helo as a homage to old time sims. You need content creation tools, asset pipeline and a stable platform (one that doesn't change too much) and two years ago it was very much a work in progress. Well we didn't exactly get all those either as it happened but you pick your battles with the intelligence you have in hand.
We use a deferred rendering engine which at the time we began was pretty new tech, only Arma II was using one and now Battlefield 3 and Unity engines have adopted it as an alternative to forward rendering. I like being able to throw dozens of light sources into a scene and see individual tracers light up trees as they pass. And I don't have to think about it.
With a game engine you tend to get scene management, audio, physics, asset management and a whole bunch of things you probably don't think about.
Yes it would have been great to use but it didn't happen. Certainly the next project requires a similar engine and we'll have a solid system of vehicle black boxes and campaign logic we can migrate.
Combat-Helo apparently looks better than the actual mission simulators used for training pilots, I can't verify this since I'm not allowed to play with those toys but that's what I was told by someone who spent three hours in one recently. I guess they don't go in for HDR and godrays in their mission simulators though
From what I've seen, although they may have improved somewhat since then, real Apache simulator visuals are pretty much flat-shaded graphics more akin to DI's Apache - or perhaps textured, but more like a smoother and less-pixelated Jane's AH-64D Longbow.
Thank you for the detailed response Flexman. Always a pleasure reading the reasoning and explanations provided by you guys. I just realized that a quick search would have revealed some of that info, but I was too eager to post, sorry.
I wouldn't even need huge maps. As an extreme case, Pacific Fighters always sooner or later became a matter of sit and wait with time warp on. Just add some additional bends to the river beds to keep us occupied (fractals anyone?). As long as the map allows some freedom in choosing your approach to the target area, while still being fun flying both directions, things will be fine I think. And a dynamic environment should keep us alert even on the way back to the base. So enough of me pouring my opinion onto you guys. It's just that we have been starved of good games and sims lately.
haha, that video definitely brings back memories. Because of it I've been judging helicopter sims based on whether they include collidable power-lines for years. Glad to see there are powerlines added to Combat Helo.
Oh lord! Firebirds strikes again...if you all only knew how much that movie is hated and despised by the Apache community. Ever since it came out, that movie has been looked upon as an embarressment and a ridiculous attempt to copy Top Gun. There is almost absolutely nothing in that movie that represents the real-life aircraft, the pilots, or tactics.
I curse you Flyboy!
Afterburners are for wussies...hang around the battlefield and dodge tracers like a man.
As a player and a developer I generally prefer smaller, more detailed environments. Any increase in scale almost always results in a decrease in detail. Outerra with it's fractal detailing and procedural rendering methods (at this point) seems to defy that rule. Not much is known about it's limitations, they usually aren't something engine developers advertise but they are certainly going to be a factor. Potential is definitely there but I'd like to see some media of it doing more than monotone grass + trees. A demo would be nice too.
Cheers
Judge, jury and executioner of Tricubic's art department.
Oh lord! Firebirds strikes again...if you all only knew how much that movie is hated and despised by the Apache community. Ever since it came out, that movie has been looked upon as an embarressment and a ridiculous attempt to copy Top Gun. There is almost absolutely nothing in that movie that represents the real-life aircraft, the pilots, or tactics.
I curse you Flyboy!
Well, I'd like to think of myself as being a part of the 'Apache community', and with a high attention to detail. OK so the movie isn't exactly accurate, but it's a movie. It is what it is. Frankly, I've taken inspiration from all and any entertainment and media with the AH-64 in it - whether it be an arcadey PC sim, rubbish console game, boring book, cheesy film or anything else I could get my hands on. I've found that ALL have something in which starts me thinking.
I partially meant my comment as a joke...it is after all a movie from two decades ago.
Nowadays, thankfully, most movies that attempt to portray the military try very hard to do them justice. Black Hawk Down, Tears of the Sun, and even Battle Los Angeles are some of the better ones. I'm not saying those movies are the most technically/tactically accurate, but they do portray soldiers (and marines, airmen, and sailors) as what they are: professionals. Not the Tom Cruise-style cowboys or Rambo-killing-machines from the 80's. OK, getting off my
Afterburners are for wussies...hang around the battlefield and dodge tracers like a man.
I maybe missing something obvious but I can't see anything really new in that video. It's the same monotone green terrain and instantiated pine tree we've seen for the past 18 months.
What I would like to see are: Rivers (with tributaries), sat map underlays, climate variation, vegetation variation + shadows, height map manipulation tools, terrain erosion/modification tools, object placement tools, controllable/fast veg/obj distribution tools, terrain tiling options, masks etc.
In terms of terrain scalability and low level detail it's clearly very advanced but until it's a usable tool it's really just a tech-demo (of which there are hundreds to get your knickers in a twist about).
Cheers
Judge, jury and executioner of Tricubic's art department.
I was thinking about this last night, and I am by no means an expert of any kind with regards to game development or the Outerra engine, so please forgive me if I come across as ill-informed on certain subject matter to follow...
From what I gather, Outerra aims to map-out the world. As far as I could see, this would only ever constitute a generic land/sea layout with generic detail across all 'continents'. It is also a streaming-type engine? Therefore can only display what it has been fed prior to execution and is not really mod-able on a game level.
For a sim such as Combat-Helo, or indeed a sim/game of any type, what would need to be done is have a portion of the world selected - and a very small portion at that - so that a developer could essentially 'close' it off (i.e. no streaming) and just have a chunk of manageable game world in which to populate with trees, objects, etc. I mean you could never make a reasonable game which has the whole world being streamed, especially for a helicopter - you'd never travel all of it! And who could be ars*d to sit there and place objects on that scale? I'm sure Flexman and AD have got repetitive strain injury just from populating the Combat-Helo-sized terrain.
Unless a lot of ground-level work is done, right now all Outerra is good for is whacking in an aircraft and flying that about over the terrain. I can't really see how any gameplay elements could be feasibly incorporated into it, as it currently stands.
As an example of how the Outerra method of streaming is NOT a good idea for low-level, ground-intensive games you only have to look at the NovaLogic Comanche and Delta Force series. Technically, these game worlds are streamed and appear limitless, but as soon as you leave the immediate mission areas, all you see is unpolulated terrain and sometimes some scattered trees. And this is on a level of only a few thousand square meters - not the entire world. You need a beast PC just for Outerra now, imagine what it would be like if you actually populated just some of the world in which it strives to recreate!
Yes, agree that Outerra engine is not really according with the current standard of game and level design, except for specific games, like simulation (jet, space or submarine why not) or MMO for exemple.
Agree too with AD, game engine without tools is just a tech demo.
just a precision, i am not a fan boy, it's only my 2 knickers. :P
My goal is just to share some informations with recent links, my english is too bad for a solid argumentation.
For exemple, the "beast" PC is not really a Shreck :
Well as I said before, I think it's superb work and I would have licensed it early on if it was available. But it wasn't.
The level of detail thing is a bit of red herring, that's down to implementation and content creation. Nothing to say you can't black box regions and populate it with entities from a streamed database which is pretty much what Flyboy said. Or you can generate detail procedurally. X-Plane 10 is going further down the procedural content route with whole cities and roads.
I'm currently studying Patrick Cozzi's excellent book on Virtual Globe design (hi if you're reading this Patrick/Kevin) which references Outerra and methods employed as well as their own extensive work in the field. It's something I want to do in the future if allowed.
It's easy to forget that it takes more than having a plane flying around to make a simulation (to quote Scott *sorry*). You need AI, events, campaign metrics, device configuration, weapon physics, all the dedicated logic for avionics (major work in itself) and something entertaining to do with it all. My biggest concern right now is making this stuff engaging. Once that's done its time to think about scalability if needed.
The other important factor to ANY terrain engine, and one of the reasons I went with Leadwerks, was performance. Performance is important to feel. And a low level terrain engine has to be able to deliver a good sensation of flight, I'm spoiled with frame rates close to 50-60fps since flying a helo at low speeds can be a bit twitchy. Coming in to land you appreciate how sensitive it can be. One of the things that can kill performance is having to re-render portions of the same detailed environment. If you have multiple optical sensors too, you'll need to re-render portions of same world for different viewports. All that eats into frame time if you need to re-do scene culling and lighting passes too. OK so the terrain is not the typical sat-map tiles you've seen but the plus side is that it can throw out a quarter of a million triangles at near 200 frames per second with nothing else to chew on. So it becomes a non-issue meaning we have time to add hundreds of thousands of trees with physics and hero ship quality helicopters and still have time to do HDR, Bloom and dof post effects. The downside is that it's not equally detailed and the scales are not what you might be used to with other sims. Still way bigger than Apache Air Assault and Novalogic areas and they were good solid playable games.
Personally I'm a bit tired of next-gen brown, green would be quite a refreshing change.
Unless a lot of ground-level work is done, right now all Outerra is good for is whacking in an aircraft and flying that about over the terrain. I can't really see how any gameplay elements could be feasibly incorporated into it, as it currently stands.
That's my opinion also. It's one thing to have a massive terrain, something else altogether to have it appear populated and alive. It will always take huge numbers of man hours to do realistic looking terrain and increasing the size of the terrain to encompass the globe just magnifies the need for man hours. For a heli sim where you spend the bulk of your time down at tree top level less is most definitely more.
Cheers
Judge, jury and executioner of Tricubic's art department.
Unless a lot of ground-level work is done, right now all Outerra is good for is whacking in an aircraft and flying that about over the terrain. I can't really see how any gameplay elements could be feasibly incorporated into it, as it currently stands.
That's my opinion also. It's one thing to have a massive terrain, something else altogether to have it appear populated and alive. It will always take huge numbers of man hours to do realistic looking terrain and increasing the size of the terrain to encompass the globe just magnifies the need for man hours. For a heli sim where you spend the bulk of your time down at tree top level less is most definitely more.
Cheers
It's not a game engine. It's a geographic rendering technology (calling it a terrain engine seems like an understatement), you can add it to whatever you need. I'd kill for some extra z-buffer precision and procedural grass (well not literally but you get what I mean).
Just because it's capable of rendering a planet doesn't mean you have to. I'm sure you can just supply limited data to suit any requirement. Ask Brano, he's a really nice guy.
There's other technologies for this which are available for licensing, however they don't like military related projects and I wanted it for a S&R platform. To me it's just software, a means to an end, right tool for the right job. Simple as that.
As for populating terrains to really high degrees of detail, it can be done, no question.
I have never called it a game engine. Infact I have been saying it is exactly the opposite of that. Calling it a terrain engine seems to be more an overstatement rather an understatement in my opinion as atleast 60% of the features needed to display an accurate visual representation of a slice of terrain haven't been demonstrated.
Just because it can render perfectly at eye level doesn't mean it's the correct tool for the job. If you move away from it's ability to render planet size terrains then all you're left with (at this point) is pretty displacement maps and z buffer accuracy. Two aspects that could be code from scratch much faster than the time it would take to code the plethora of other components, features and tools needed to visualize and develop terrain on top of outerra.
Lots of things can be done, but as we know, due to constraints they don't get done, which leaves huge manual tasks in the hands of content creators and ultimately the quality, accuracy and realism of the terrain suffers.
Cheers
Judge, jury and executioner of Tricubic's art department.
From what I gather, Outerra aims to map-out the world. As far as I could see, this would only ever constitute a generic land/sea layout with generic detail across all 'continents'. It is also a streaming-type engine? Therefore can only display what it has been fed prior to execution and is not really mod-able on a game level.
For a sim such as Combat-Helo, or indeed a sim/game of any type, what would need to be done is have a portion of the world selected - and a very small portion at that - so that a developer could essentially 'close' it off (i.e. no streaming) and just have a chunk of manageable game world in which to populate with trees, objects, etc. I mean you could never make a reasonable game which has the whole world being streamed, especially for a helicopter - you'd never travel all of it! And who could be ars*d to sit there and place objects on that scale? I'm sure Flexman and AD have got repetitive strain injury just from populating the Combat-Helo-sized terrain.
I'm not quite sure what you mean by a "streaming-type engine" here. For one thing, the engine can progressively download and cache the terrain data, depending on the level of detail necessary for given view. This is in order not to have to download the entire dataset for earth, which is ~13GB for elevation data and will likely grow to 20GB with auxiliary datasets. However, when all you need is a local 100km x 100km area, though with unconstrained full detail vision towards the horizon at its boundaries, you'll need just few hundreds of MB.
For another thing, the engine is procedural, so you absolutely don't have to populate it by trees and other stuff by yourself. It currently uses a model that computes the probabilities of vegetation growth using local terrain topology, combined with fractal algorithms to emulate the natural growth patterns. We are also going to use data that define rough maps of world forest coverage, in combination with the existing techniques, to get real-looking forests.
Ultimately, using coarse world land class data and ecotype fractal-driven models we aim to get earth in its pristine natural state, i.e. without the effect of human civilization. Over that then comes a civilization layer which is mostly vector based - spline based roads, polygon areas that alter the naturally occurring terrain to make fields and pastures, urban areas that will be populated by a city generator, and so on. Much of that data can be imported - for example using OSM data. The civilization layer is seamlessly merged into the procedural terrain so that the transitions are fractal-smooth, and the approach even allows for variable decay of the civ layer back into the natural state.
Returning back to the original issue - when you leave the designed mission area in Outerra, it still can download the additional terrain data for you automatically on the background, you'll still get the finest terrain resolution with all the vegetation, a believable natural world - that's independent of what your game designer does. You'll miss only the civilization or the customizations that were added by the game/sim. Outerra changes the way how the game world is being made - developer won't get an empty box to import and model everything in it, he'll get a whole planet that's ready to wander on, without him even starting to do anything.
Of course, as you can surely imagine, it's all a huge amount of work, and so everything can come in only gradually. There's a lot of effort to make things smooth and optimized for the modern graphics hardware, which is what made this approach feasible.