nbohr1more

To add to the discussion: ATI Catalyst drivers past 10.11 no longer appear to properly disable Catalyst AI. In general there are two workarounds: 1) Use a 3rd party tool like Radeon Pro to disable it. 2) Rename the executable (for Doom 3 this would be darkathena.exe). For option 2 to work with The Dark Mod, you would need to place the map pk4 into The Dark Mod base and open it via the "map" command in the console. Tracker 2725 was opened because the ability to specify a different executable name is broken in the Windows TDM_Launcher: http://bugs.angua.at/view.php?id=2725 The original Topic: http://forums.thedarkmod.com/topic/12398-disabling-catalyst-ai/

That first concept menu looks familiar... I saw this over at Doom3world in the "Doom 3 Can Do It Too" project thread...

Looking good.

A beauty as always Johannes !!! (I'm trying to keep my spirits up and hoping my hardware can handle it )

Final Fantasy is widely regarded to be a mixed-bag as of the latest releases. Square Enix seems to be coasting on the momentum of the concept that Final Fantasy is an invincible money maker. If they retreat to improve FF, it's hard to say whether the efforts will payoff. They already pour tons of money into the production quality, the problem is that they have all but forgotten it's a game and care more for the cut-scenes than the story tree or player choice and consequence. If more budget is afforded to FF, I can almost guarantee that they will just make shinier more expensive cut-scenes and have no clue that this won't help. If they were smart, they would take a hard look at things like Heavy Rain and LA Noire and see where the future of interactive RPG gameplay is headed. I suppose they took a look at World of Warcraft and tried to follow that trend but MMORPG is too big a mountain to climb. FF was never deep enough to sustain the vast hierarchy of a hardcore RTS and thus does not have a fully tuned system to translate they way Warcraft turned into WoW. If they wanna go WoW they'd better explore RTS first and not in the paltry manner that was done for their shovelware DS title. Again, a move to MMORPG is a betrayal of the core single-player RPG experience that the "FF series was once acclaimed for" so modern advances in single-player RPG narrative and play are a better route to take. (But rat-bastard corporate types probably have no concern about keeping a game franchise in a recognizable form for it's fan-base... cough TDS ?)

This contest rocks!

Very good. Can't wait to see your mission Jesps

Lights per scene? I though that r_showLightCount shows light count "per surface". The old belief was that you should not exceed 6 lights per surface but fog, blendlight, and ambient light counts can be much higher without too much detriment. Shadows are the primary problem so light overlaps for non-shadowed lights are much more forgiving. That said, every light overlap duplicates verts so you can get geometry bound with pretty blocky scenes if you go crazy with the lights. ~~~~~~ After studying the topic over and over again, this is the ideal setup (it's a pain in the ass though). 1) Setup you primary static shadowed lights 2) Create small rectangular "overbright" ambient light volumes close to surfaces for decorative lighting (using custom textures where necessary) 3) Create free-standing ambient volumes (or even regular no-shadow volumes) that don't touch the surfaces (levitating?) for subtle lighting effects on dynamic objects (AI, etc). 4) Cut your geometry to keep light counts down Step 2 is essentially like putting a "blend add" stage in your material with a small light map texture but it's a little more flexible because you can move the volume relative to the surface. Step 3, light volumes overlapping does not have a severe performance impact except when these overlaps hit geometry. Just be careful because if an AI passes through a boat-load of overlapping volumes you will feel those FPS melt away... So that's the best o' the best "manual" way if you are a monk with a 1000 yrs to light a map. The attached image shows: A cave wall ortho view (BLACK) A Projected Light shadowed (BLUE) A decorative Ambient (RED) A floating ambient (GREEN) Illustrating the different placement scenarios above.

This is probably already a known document but I thought I would link it here just in case: http://software.intel.com/en-us/articles/optimizing-the-rendering-pipeline-of-animated-models-using-the-intel-streaming-simd-extensions/ If I'm not mistaken, all the animation code is in the SDK? ~~~~ Tangent: This may be a little cheeky of me but I brought up the concept of using animated normal maps for facial animation in place of bone based animations. If I understand that flow-chart correctly, there are early-out branches when no polygons are changed. Therefore, my theory appears to be sound. It would offer a performance savings. I must stress, if I hadn't already, that this is of purely academic interest. I have no demand for Springheel or any other modeler to build a bazillion high-poly head models and bake a bazillion normal maps to implement this.

Patches broken per light.

Add me to your list if you wish. I think I can squeeze more testing time than I had for RTTCv2 (that was a unexpectedly busy period...)

I think it was cool that Splash Damage went ahead and expanded Megatexture for it's version of Id Tech 4 but I think this has been an ongoing coordination with Id. Kinda a cross-promotion between the two. That said, returning to baked lighting for megatexture games which both this game and RAGE seem to have done is definitely a bit of a disappointment. If the megatextures were only used for texture capacity problems and the light model were left dynamic I would be a little more excited. The game seems to have it's appeals to fans of Splash Damage's match system from ETQW but it doesn't hit my "Big dumb shoot-fest" sweet spot, nor is it a subtle RPG-esque experience. Instead we have something that appears to be a niche evolution for Q3 DM > Wolfenstein ET > ETQW in terms of squad and multiplayer coordination. It doesn't appeal to the Modern Warfare \ COD crowd nor the Unreal Tournament or Quake MP crowds per se. I'm sure fans of this niche are overjoyed but for the audiences it was marketed to it's probably as much a let-down as Brutal Legend was for 3rd person hack and slash fans. I am hopeful to see what Human Head does with Id Tech 4 in it's last iteration as their changes seem to be more natural and use the engine for more of it's inherent strengths. Though some of the characteristic black shading on the characters in the early screens looks like they too may be retreating from dynamic lighting and baking things in The game-play as described in Prey 2 at least seems more compelling to me verses what we are offered in BRINK. But, as I said, BRINK's niche is not for me and if they were more forthcoming about this then it would've saved them some blow-back from folks expecting something else.

New rule: Springheel must create new missions annually! (New missions = new assets ) @Siyah You always post such great screens, I hope to be playing one of these one day ...

I agree with Rich. Focusing on a limited set of variables and making them work well together is a key aspect of good game design. Leaving the set of play variables open-ended is bound to give mappers problems establishing an intended narration and play focus. To exaggerate: Why would you sneak past 3 guards to capture a trinket when you can fly around the map all day in a hang-glider? If you offer too many choices the players will not follow through in the intended manner and will likely get stuck in places where the mission's prior activities are supposed to culminate at a puzzle or the end of a quest. I suppose, as a toolset provider, that you could argue that "mappers should be given as many choices as possible, it's up to them to fashion these limits". I would suggest that having a great core set of play-tested elements lends a quality and consistency for mappers who use the defacto stuff.

All entrants are eligible to revise their missions until June 10th.

No problem. Shall we tentatively set the extension to June 10th?

Doom 3 loads all the textures for the whole level at once. Because of this, most of the initial views you would see are not going to be paged out to "whatever the PCI-E equivalent of AGP Aperture" (PCI_E Aperture?). Normally engines would only load a small fraction of the map into vram then page in and out when you've passed certain distance markers. If you have filled-up your vram in those types of engines you have no place to swap so therefore the game becomes very laggy.

What happens if you split those blocks into their own FS? Does it look similar or worse?

The intended settings for the mod are to use compression where available. Not using compression may decrease your load times but, as you can see, the maps are way up there in vram consumption when compression is not used. Both RTTC and FAFS are relatively small missions, when larger missions appear you may get malloc errors. It's cool that you've got the hardware to handle this but I suspect you would not see any appreciable quality loss for enabling compression. (Especially since 99.9999% of the normal maps aren't compressed anyway...)

OpenGL can have less draw call overhead than DX is many cases. It is entirely possible that poorly optimized maps will perform worse than what you've experienced. Oddly... part of the reason that Deferred engines like UE3 perform so well is that they have more scene management overhead to prevent overdraw. In cases where you hit high vram usage in these engines you are pushing the edge where this overhead starts to be more of a burden. Though the tradeoff is worth it because UE3 can make scenes that look like the equivalent Doom 3 scenes with much less ram. Some UE3 detractors would say it isn't worth it due to the buggy way that deferred render engines deal with AA and filtering... I am a little surprised by your vram numbers though... I know that in v1.04 some folks found that texture compression was disabled... are your DoomConfig.cfg settings as listed below? image_usePrecompressedTextures "1" image_useNormalCompression "2" image_useAllFormats "1" image_useCompression "1" image_preload "1"

YEE HAWW!!!

Article has been posted: http://www.moddb.com/mods/the-dark-mod/news/winter-harvest-released It will take a couple of days to authorize. Hopefully AluminumHaste will get a chance to collect some pics to accompany it shortly.

Definitely caveats are in play here... I always thought blendLight makes better looking fog but that is kinda like saying lakes get you wetter than swimming pools. They both control the same volume, texture and, opacity values... fogs just use a "distance to player" value. The real mind-warp is to consider the effect of a falloff image on a fog. So it can be thicker by height and thicker by distance to player... Katsbits is applying their fog over a large parallel light so their parameters might not be kosher for smaller light sources. So the gains there are still questionable from a performance perspective. The artistic options are definitely interesting though... Those Phobos guys must've beat their heads against rocks trying to make their maps look good and, missing an ARB shader guru, came up with this creative use for fogs.

I was looking at the latest Phobos mod screens today and noticed the use of a blue fog to simulate atmosphere haze in a very deep shaft. I was admiring some of the scenes for the pure artistry that was used to shape the scenes so that non-overlapping lights look natural. After wandering the web awhile I went back to searching for Doom 3 lighting topics and re-read Katsbits "Outdoor lighting article" and was struck by another concept that previously went under my radar: So what they are advising is to use a very very thin fog to achieve a similar light softening that the Bloom mods or Post Processing achieve. I wasn't sure about this advice but went back to admiring another Phobos screen where a commenter said: "Nice Bloom" To which the Phobos developer retorted "We don't use Bloom" Then another thought clicked... I recalled somebody puzzling over why Lunaran used a Fog volume for the ceiling light in Strombine. It all makes sense now. Fogs can blur the edges of where light is hitting and make it appear to scatter in a similar trick as Bloom or even radiosity. Glares on light sources also have some of this spreading effect. So... I know that overall fogs seem to have a performance detriment... Does having lots of small localized fog volumes have significant performance impact? How does this compare to the enabling Post Processing shader? The Phobos team seems to have mastered the artistic use of fogs, I wonder how their maps perform...

The vertical contest was half the reason I pushed for "someone" to update Moddb (little did I know that "someone" would be me ). I thought it was a crime that the vertical contest wasn't promoted there. Baddcog's mission belongs there as well as a whole slew of the initial batch of missions. (There is still no consensus about whether all missions should be up there for an alternate download location...) Baddcog's say is good enough for me, I will have a Winter Harvest article up soon.