Flexman

Thank you sir.

Same goes for "Wings of Prey". I believe EA are considering pursing the same policy as Ubisoft for more titles. It's also why I confined the XBOX 360 to a cupboard and bought a PS3 - to explain, if you upgrade the 360 HD, or as in my case sent the console back for repair, a different console was returned. All purchased games then required a net connection or they would only run in demo mode, or not at all. As this is not physically convenient, away it went. Besides it was too loud and had been returned so many times the kids were calling it the "Fed-XBOX". I won't pirate, I just don't buy these kinds of games anymore and search every box for that "Internet Connection Required" stamp. Plenty of other games to play and I can live without the hassle. Oddly, I don't get on with Steam either but for different reasons. I bought the new Empire Total War/PC on DVD published by SEGA. When Amazon delivered it I proceeded to install it. Much to my annoyance I couldn't play it, it wouldn't let me play for another 24 hours until Steam decided the game was 'live'. Not having much time to play games I'd grab what game time I could, three times I tried to play Empire Total War over the course of the next couple of weeks. Each time I was required to download a patch, the duration of which sucked up a good percentage of that time. A boxed game I've hardly ever played. Asassins Creed II is a great game though.

Looking at those links makes me want to weep a little. We're still trying to populate fields and scrub, after a "few" hours so far we have something akin to Farming Simulator :/ I think we needs to seriously work on some new veg textures, some new veg models (as those are too small and need filling out) and a bit more tree variety. We have 3 veg slots left and we might use those for animals and civilians. It's pretty nice to fly over though. Suggestions on lighting, painting welcome.

Yes, Michael. That would be a great help. We were wanting just such a thing yesterday. For that very reason. While putting together this game, using the current tools we keep hitting the same things that get in the way of producing content. I've tried to put these in order. Some of them would seem (at first glance) fairly trivial and help a lot, but I'm a bit clueless as to how difficult these actually are. Vegetation Fill veg by paint layer' - "I want all areas that are painted with layer 1 (grass) to have this vegetation (with the hight/slope constraints)" as outlined by MB in the above post (funny we hit the same thing, says a lot). 'Use constraints when placing veg' - If I set a veg layer to 'slope 60' 'height 200', and manually place, it'll place the veg at any slope or height. This is a problem when painting hills because they pop out of the steep sides of the hill instead of just on top. 'Clear all vegetation from this layer' - i.e "I dont want my dirt track layer to have any vegetation on." Top down billboard for trees. Paint Brushes. More brush shapes would be great for sculpting/painting/veg placing - Square (for example for fields), hard pencil, spray can. Just a simple option to choose between square and circle would go a long long way. Roadside scrub, narrow fields and gardens. Line sculpting/painting/veg placing. LMB to start, hold SHIFT, LMB again and Terrain sculpt, or texture layer paint, places veg in a line. Number of Paint layers. 6 or 7 would offer much greater variation. If more layers are impossible how about the ability to effect the brightness/saturation/contrast of the brush. This would allow the painting of shadows under trees at great distances, have terrain selectively shaded by cloud cover. 'Export the true colour map by visible layers' - i.e inspect pixel X,Y and build each channel (RGBA) of the DDS with the correct levels rather than filling the entire channel - Much easier to manually edit. Terrain Road model export support - Allow us to export/import the generated road as a model So it doesn't have to auto-generated at runtime; Too slow at 4096x4096. Or compile the road section generation into the engine for speed. River support - A tool similar to the road tool that can dig a slit trench from A - B and lay a flat road-like water surface along the path. Options to adjust width would be helpful. That would greatly improve content creation IMO after many many hours (man days even) of use. Just to illustrate. Lets say you were making a level with paddy fields, lots of small channels, fields in a line. At the moment we're having to do this...create a layers for blending onto the heightmap. Which after trial and error, swapping between Photoshop, exporting and loading in the Editor produces results. Then those paddy fields require hand painting with various veg layers, with a round tool (cries). The river has also been carved out by hand in a similar manner to the fields, the river is a 3D object (in sections). The dirt track running east/west, a texture layer. Would be much easier with a line tool.

As our helicopter can be called upon to fly missions day or night, having instruments you can read in the dark or see in low light considtions, at a glance was done just like our MFD screens, using the fullbright shader and setting the light map into texture stage 2 texture0="abstract::apache_pit_instr_01_skin.dds" texture2="abstract::apache_pit_instr_01_light_skin.dds" shader="abstract::mesh_diffuse.vert","abstract::mesh_diffuse_fullbright_glow.frag" When the instrument needs to light up, I'll paint the switch with the "lit" version of it's material. And normal version of the material to turn it off. Here's a screen from my blog to illustrate. We started experimenting with vegetation in the editor on our production terrain (digital elevation data of a region in Afghanistan). It's bloody hard work getting anything done. At our map size the auto populate feature just doesn't work, often crashes. In fact, at the largest map size most things will upset it at some point or another. It's been a really frustrating experience. Had a little experiment with stretching a texture across map but the amount of bluring at 4096x4096 @10m per tile didn't seem to add much. Not at mid to low altitudes anyway. I'm sure it's great for making smaller maps. Maybe you need to have a real talent for blending. I dunno. Two more texture stages would be great, a zoned vegetation populate function that worked would be worshipped upon the alter of LE game development. I'll even supply the gold idol (but only if you throw me the whip). Everything is so far ticking along nicely apart from the editor issues and various odd glitches and artifacts. When you climb into the chopper, look around, the sound of two engines thumping through the speakers, head tracker software and shadows of each rotor blade dancing all over the cockpit, on a huge screen its pretty ***** cool. Using Leadwerks Engine has been very love-hate but I'm loving the results (most of the time). Seem to get a glitch in distance fog at the terrain horizon, small gaps, or short dark lines. Why does the editor crash using the vegetation auto-populate feature on the largest map size? Is it just trying to allocate huge amounts of memory?

Micheal, can you explain a little about the blending a sat image? I'm so heads down in other matters. We have the 5 textures already allocated (could really use two more for greens). Is it possible to use an additional map to colour everything? If so that would be of great benefit and I'd would kick myself for having missed this. *** edit *** Ah, yes I see. We sacrafice one layer for one huge overlay. That might work to give it some interesting variation. Thanks for the heads up. Still could use an extra layer or two :/ but we work with what we have.

Out of context it does sound like a contradiction but let me explain. Not all controls are active any any one time, the bulk of that is 5 MFDs each with 24 buttons dotted around them, the screens are like your local ATM, with arrows pointing to the buttons that only do something for any given page it's showing. You might want to launch a missile using one mode, or have it scream high into the air and plough down into the roof of the target. The purist can push all the correct buttons to bring up a weapons page, select the LOAL or LOBOL launch mode, lase the target and launch, the helicopter will even select the next target for you so you can keep shooting till dry. Or you can hit the right joypad bumper to make BIG the square in your HUD and shoot. The joypad sends the same message to the avionics class in one click as does navigating the purist menu system. The comms panel looks daunting, most are just volume knobs, like a mixing desk. Adjusts the volume of 5 radio sets. Not necessary to gameplay but add flavour to the environment and easy to include. We have speech from ground forces that periodically request support, you can turn them off, not very nice but hey. The trick is cramming everything you need into a joypad, this helicopter automates a lot of things. Pylon select, radar mode, TV, shoot missile, guns, launch mode, autohover, autopilot, master mode cycle (arranges intruments and display into preconfigured states). Up down, left right and a view select. That's all you need to play. Then you can work deeper and learn how all these things work together as you get better.

The CH53 sure is a big helicopter that needs a lot of looking after. I'd like to apologise in advance for using the word "Helo", I know some don't like it, "The Army have choppers, not helos". But it's not as catchy at retail :/ Of course I'm making an assumption there sorry, we're you Navy? Coast Guard? E suggests Marines. We got told that certain things are off limits, countermeasures and *unspecified classified* systems. The rules on taking photographs seem to vary from unit to unit and who you were :/ But the best source of data came from the MOD where you can take a virtual tour if you have Apple Quicktime installed.... http://www.army.opticalfx.co.uk/popup/tours/hi/3.htm You can probably match up some of the textures in our cockpit with that. My previous helo work is all PC based. I was involved with Empire Interactive's Enemy Engaged Apache Havoc and its sequel Comanche Hokum, in 1996 a mission planning tool for Jane's Longbow 2. Before that, Microprose UK. So a fair amount of experience with the PC sim market and helicopters in particular, no reason to that, just turned out that way. It's the devil you know I guess. We're trying to re-invigorate the market of fun combat sims that were popular back in the 90s when they didn't take weeks to learn or years to make. Those games had a compelling game element that kept you coming back. Today they originate from eastern Europe where they have a heavy engineering background commendable for their accuracy but all very dry and, have to say it, not that entertaining. This is purely an entertainment product that also happens to be a sim, not military spec but light to mid-weight, while it has sub-modes for the fire control radar and authentic HUD system, I develop with an XBOX360 joypad as well as a full HOTAS, it can be picked up and played without much prior knowledge. Leadwerks Engine is something of an experiment, while written for making FPS games, it provides enough low level detail for helicopter and heavy armor games. Helos don't fly very fast and often used in a ground support role, CAS or saving lives. But thanks for all the kind remarks so far. It's all very exciting to see bit come together. I can't stress enough about the importance of planning what you want and don't get sidetracked by feature creep. Just say "no" to gravy.

Here's a short video of it in action with TrackIR (no chance of keeping the camera still there). http://www.youtube.com/watch?v=uLvCX9XXZSE Need to finish off all the wiring of the switches to their message IDs. It's a long process. Our production terrain map is in progress, although we have some issues to work out. The base terrain is in place ready to start hanging details from it. The view distance will be pulled back a bit in game but this is an editor grab...

Thanks guys, I'll post a more detailed explanation of some of the other artifacts and issues we're having a certain cam ranges, I just need to compile my data, screen-grabs and organise it. Some are just strange lighting artifacts that manifest on terrain or on some (not all) 3D objects. I'll post again in this thread when I've got all that sorted.

Hi gang, Here's a head scratcher. We're making a large map based on a real region. 4096x4096 @10m per tile. This is the maximum option offered by the editor. At unit per meter that's close to 40x40km of space perfect for our lil-old tank busting helicopter game. Placing objects near the outer regions of the map in the edit, and in game result a lot of really ugly z-fighting artifacts. It's really bad in the cockpit too where there's surfaces close to each other. Now the cockpit situation I can mitigate as it needs to go into it's own world anyway. I presume this is due to floating-point errors at distances from the world-origin. So we reduced the terrain down to approx 20x20km with 2048x2048 @ 10 meters per tile which helped a little but it's still not great. Is it feasible do you think to translate the terrain and scene root to always maintain the camera at 0,0,0 ? Cockpit parts flickering like crazy all over. Objects outside too.

We have a current need to populate small regions from GIS sources with vegetation maps. The current editor runs out of memory and falls over if you try and auto-populate the largest possible map size. Not that we want to auto-populate such a large map but there are plenty of compounds, rivers and fields with vegetation. We would love to be able to generate vegetation in code to make an importer. Any information on storage format for this would be cool.

We've been making good progress with complex entities in Leadwerks engine. I thought I'd share a few lesssons learned which may not be best practice but have been useful all the same. The most complex entity of the game is the cockpit area of the two-seater combat helicopter. Our model which has working avionics and rudimentary flight model features exactly 340 user click-able controls in a model that has close to 34k polygons and other shapes. Using CameraPick() seems easy but often, child objects in a complex mesh are hidden by others. We elimitated any problems with camera picking by naming all interactive objects with a simple keyword embedded in the object hierarchy. A small section of the list would read... cpg_comm_switch_hfon cpg_comm_switch_ics cpg_comm_switch_ident cpg_comm_switch_iff cpg_comm_switch_master And passing a filter function in CameraPick(...) to only return objects with "switch" returns a good result everytime. Function PickFilter:Int(ent:Tentity) If (ent.GetKey("name").Contains("switch")) Return 1; End If Return 0; End Function This results in perfect picking of all switches within a busy model. Another thing we did to improve performance by a few percent was organise the hierarchy of the model so areas not visible to the player were located under a parent object. And we would "show()" and "hide()" that parent accordingly. Some thoughtful organisation of detailed assets can yield benefits in terms of ease of programming and performance. I hope this might be of some use to others.

That looks *really useful*. Keep us appraised, we could certainly use it.

Josh, I really like the current logo and gears. I've been graciously supplied with a suitable version of the sites current logo which I think is quite stylish and perfect for the job. Like programming, art is never finished, only abandoned. Take care out there at the GDC.

Nothing obvious comes to mind. Dunno but you might want to look into SetWorldSize() here's the wiki entry... SetWorldSize wiki Your video suggests it's some odd lighting/shadow quirk. But try SetWorldSize something large.

There are other issues though. Josh can't just write the engine for XBOX or PS3. Microsoft won't let you near it unless you can demonstrate x,y,z and jump through hoops backwards. Red tape. If the engine could leverage a reputation of having some AAA titles behind it then maybe it might get some wheels turning. Or a stonking huge bribe.

This is what our animator wanted. He was trying to explain to me that having lots of animations in each character was somewhat wasteful and it would be much better to have lots of different characters you apply an animation to. It seems obvious now. We're starting to look at infantry and various various civilian characters which would benefit from the same animations applied to a range of slightly different models.

Could we have vectorised or high resolution LE logo for splash screens and videos?

How did you implement the cables from the electricity pylons? Are they a 3D model you load in or do you have some neat script that generates them? I 5* your video. Gate to the Minds Eye stuff but real-time.

If you mean floating text, over an objects postion? CameraUnproject(), in bmax and example might be... Local tpos:TVec3 = CameraUnproject(mycamera, entity.position.Plus(Vec3(0, 6, 0))) ; ' ADD AN OFFSET ABOVE DrawText("I can see my house from here", tpos.x, tpos.y) ;

I have a character I want to make invisible but retain it's shadow. And I need to make him invisible without altering other characters that use the same model. Using the "invsible.mat" on a model changes all instances of that character. But I don't what it rendered at all except for the shadow. Any pointers?

Well it's good to know about the INI files. One way or another. So from "LODDistance=120" to require("scripts/class") local class=CreateClass(...) function class:CreateObject(model) model:SetLODDistance(120,0) end Would be handy of the LE Editor could add a LUA file to entities that don't have one. And the option to stamp in some default values like mass, collision type, LOD. Then again, seems a bit lazy. Cheers and thank you.

I have a question. Are INI files going to be phased out completely? If functionality from these files are being steadily dropped requiring a LUA script for every object, I'd rather get started sooner rather than later. Having a global SetLODDistance override would be very handy for sniper scopes, TV cameras, that sort of thing.

I think that's brilliant. The social aspect is mind boggling. Games are the only medium which deny you content if you're not good enough. This would allow anyone to see all the cool stuff you never get to see. Compressing a frame in 1ms, fantastic. Never need to upgrade or spend half a grand every two years to play new games. To say I can see the appeal is an understatement. Buy shares now.