DM-AMC-Ars_Indus

[Swanky]

Got the map into Beta status. I'll update this post with new / final versions when needed.
Current status: Final 1.2

DOWNLOAD DM-AMC-Ars_Indus

Screenshots
Screenshots appear slightly darker than ingame for me.

Spoiler

Installation / Requirements
Done on UT v4.36, Ued 2.0
Richrig.utx (included in download)
Drop .unr file into your map folder, drop richrig.utx into your textures folder (if needed). Play.

I do recommend using OpenGL drivers, especially the community made ones, since a few textures can look ugly when using D3D renderers.

Couple of notes
The base layout is around 5 years old; I think in between the time of Peak17 and Narean, with the center room originally being created by Creavion with his map Ryigiar. I scaled the layout down, added another floor and two walkways on one side (the blue corridors behind the pillars) and had the layout intertwined. It's a rather small map which by itself is not unusual for me, but it's also a very tight map. I always wanted to do a white metal / brick industrial map, but after a few days of on/off work on the map I severely lost any interest in the map.

Fast forward to late 2018. After I saw both Revelation AND fragnbrag returning (at least briefly) to UT, I thought of doing a map. However, with just 3 weeks to go for the AMC contest, there was little hope of actually doing so. It was when fnb pointed out to me that the point of having to do a map from scratch was scrapped itself that the lightbulbs and itches to do some mapping came back on. I decided to finally finish that little layout that had sat on my harddrive for so long...

Mappingwise, the outdoors took way too long to do for what little time you actually spend there, and its lighting drove me nuts because I originally wanted the map to have pinkish lighting instead of orange. Of the originally one week that went into the project after I came back, the outdoors alone ate 3 evenings. In hindsight I should have just walled off the area even if it did give me a few nice prefabs for next time. The worst part of it was lighting, however. Since sunlight does have a distinct color, the map was too small for having sunlight with blue shadows, a light blue lighting color and a pinkish main color in the complex. I originally wanted the map to have a bit of a Mass Effect feeling with its colors, but thanks to the sunlight the map felt far too colorful and the idea had to be scrapped.

It also took a bit of time to get the bots working to get around the map and not just center in the what is now the Pulse Gun room. While they are relatively smart, due to the constraints of the pathing abilities in UT they cannot liftjump to my knowledge, giving human players a distinct advantage at getting the armor. Because of it, the upper area is slightly less populated than I would wish.

Stuff about the map itself
- is a 2 - 5 players map depending on your playstyle. 1on1 is rather slow with a lot of hide & seek. 3 to 4 players is the sweet spot while you can fit in 6 people in team games as well.
- All weapons present except for the ripper and warhead launcher
- 8 player starts present
- bots get everything; however, until I find a way to make them liftjump, humans do have an advantage at getting the body armor on the third floor
- polycounts get as high as 1000+ in the outdoor area looking towards the crane & pulse gun room but are otherwise around 400 - 600
- 2443 brushes with a ratio of 2.41:1; the map runs at constant 60FPs on my system (OGL drivers)
- it is significantly brighter than my last few maps (Selket, Peak17 and Narean); so there shouldn't be any brightness complaints this time around

Story
Being in the lightbuld service for tournament lighting, the steady decline of demand for new arenas and lightbuld replacements drove the company into insolvency.
Now being owned by Liandri, the narrow corridors and intertwined layout serves as another arena for their annual deathmatches.

To Do (not that it's going to happen)
- teach bots how to liftjump
- find another fitting, slow, mysterious, industrial music theme

Known Bugs
- The crane textures may appear wonky in D3D; I recommend using OGL Drivers

Credits
Hourences (textures)
rorshach (textures)
HortonsWho (textures taken from Ravencross)
Eli (?) - custommapmakers.org/skyboxes.php (Skybox textures)
Akuma (textures)
Creavion (original center layout)
Revelation (Sunflare actor, Footstep Manager, skybox help)

The map has released in final. This is the official download thread for the contest. Any feedback is of course appreciated.

[papercoffee]

It looks fantastic, only the Skaarj computer panel looks off and breaks the immersion, in my opinion.

Edit-----------------
So, this will also be the thread for the final release.
Linked it to the contest thread first post.

[R3plicant]

Very tidy and competent map, Swanky. I have to ask though, why are most of your textures set to High Shadow Detail? There's really no need and it only causes slow-down. You can leave it unticked for MOST of the textures and only set it to HSD for textures which are receiving important shadows. The map is compact, tidy, well designed/made and I think it's a strong contender for the AMC competition IMHO, good little map.

[Swanky]

Thanks, both of you.

@papercoffee: It's a homicide argument. Immersion can be VERY subjective. Also, replacing it with standard computers usually means the angled curves will be gone, and I don't want that.

@R3plicant: Is slowdown still that much of an issue these days? Both Revelation and I ran the map at constant 60FPS pre open-beta.

[Aldebaran]

Good map with nice layout, textures and sky - loved to play it! No issues found here. Behind the door of the small tower outside you could place a pickup

[R3plicant]

@R3plicant: Is slowdown still that much of an issue these days? Both Revelation and I ran the map at constant 60FPS pre open-beta.

If you have a new-ish PC that runs new games at 120fps (just saying: hi-spec), then sure, your map should play smooth as silk, but if anyone else still has an older PC, then it's nice if it runs the map smoothly on older hardware.

Set all your textures (except unlits, invis' of course) to low shadow detail and only set any to high detail if they are receiving important shadows then rebuild the map and look around for any visual abnormalities, correct where necessary and rebuild again. If all is ok, then save and know that your map will run smoothly on even the older spec'd machines still in use. If you REALLY want to go all out, set any textures that are not visually important to BOTH low and high shadow detail (on the same face/texture), this will make their shadow detail extremely low and will optimise the map still further.

My knowledge of advanced optimisation is admittedly limited but this I happen to know (assuming what I've read is true).

[Swanky]

For that kind of optimization most of the computers in use had to be at least 15 years old. I've optimized the high shadow detail surfaces and added a bit of lighting detail to the crane and the wall close to the bigger lift, otherwise I think I'm done for now. I'll wait for a bit longer with a new version, though.

[sektor2111]

To Do
- teach bots how to liftjump

Easy stuff... MBot does this already...

Spoiler

Code: Select all

		if ( Ow.MoveTarget != None && Mover(Ow.Base) != None 
		&& ( Mover(Ow.Base).Velocity.Z > 80 && Mover(Ow.Base).Velocity.Z < 400 )
		&& FRand() < 0.7 && !Ow.Region.Zone.bWaterZone )
		{
//			log (Self.Name@Ow.getHumanName()@jumping from lift);
//			log (Ow.getHumanName()@ has velocity z@Ow.Velocity.Z);
//			log ("Lift Velocity on Z is"@Mover(Ow.Base).Velocity.Z);
			foreach RadiusActors(class'LiftExit',Le,300,Ow.Location )
			{
				if ( Le.Location.Z > Ow.Location.Z
					&& FastTrace(Le.Location,Ow.Location+vect(0,0,12))
					&& Ow.Location.Z+150 > Le.Location.Z )
				{
					TheGoal=Le;
					break;
				}
			}
			if ( TheGoal != None )
			{
				ZPower = Ow.JumpZ;
				Ow.JumpZ = ZPower+Mover(Ow.Base).Velocity.Z+50;
				Ow.StopWaiting();
				Ow.SetBase(None);
				Ow.SpecialPause = 0;
				Ow.PlayRunning();
				Ow.MoveTarget = TheGoal;
				Ow.Destination = Ow.MoveTarget.Location;
//				log (Ow.GetHumanName()@has@TheGoal);
				Ow.BigJump(Ow.MoveTarget);
				Ow.MoveTarget = TheGoal; //Regain this after jump
				Ow.JumpZ = ZPower; //restore jump
				Ow.EAdjustJump();
				Ow.SetFall();
			}
		}

Change it with other stuff if anyone else has a better one...
Legend:
Ow means a Bot owner of this tool.
To Do Hints:
- looping through Bots in a single iterator ?
- Coders with experience in tracing, looking for a random spot around LiftExit destination and attempt to jump there if possible ?
- Use RouteCache[x] instead of iterator ?
- any other viable solution ?

[Barbie]

To Do
- teach bots how to liftjump

Easy stuff... MBot does this already...

But the formula you use in the code is only a guess where you calculate with fix values ("150") and presume that the bot can jump 150 UU regardless of gravity and lift speed. A real calculation has to differ between the mover EMoverGlideType and have to include the gravity.

Another interesting question would be: where is the point to jump off to make the highest jump? (Create a formula, derive the function and determinate its zero crossings.)

[sektor2111]

If you remove log's remarks and add more logging you'll see that Bot will need an extra-velocity else won't do any jump like human does. What I posted here it's already used for a few months and I assure you that it works. I listed hints disturbing me, it's not about velocity it's about more randomization, perhaps I'll do that but not so soon, it's not a priority - so far this is suffice for me. All those lines are based on a study of Bot code and Mover base during games, find another way if you can do it better and then I'll change my code with yours.
Player jumping from lift can accelerate depending on lift speed (lift actually do seems to assist player in this jump out) in really high spots (DM-GrendalKeep_Rv as example it's like you have jumpboots). I copied the same stuff at Bot (a bit hackish) because actually bot having normal default Z velocity will only jump like it does on the ground without to look like mover was boosting it like in Player case - Bot has 0 VELOCITY.Z on a lift-base - log this and see, and then we can speak later. The only velocity belongs to mover, Bot has nothing so it will look retarded when is trying to jump in original way. Faster movers might even kill it if I well recall primary tests, Bots accelerated slower than Lift speed because their velocity was ZERO and lift crusher was killing it in the most of cases because bot was leaving lift-base but moving Slower than lift. I know what I was doing even if it do looks hilarious.

Feel free to see first how do this work and let me know about traces... Also you can invite me in some maps in your sever and let me show you some Lift-jumping a la Myself and then... go figure why I wrote those lines for Bot... or we can fly together in Monster Gaming Server.

[Swanky]

I uploaded a new beta version. I updated some textures to be more performance friendly (high shadow detail) and fixed some lighting issues on the crane and some pipes. The skybox received some new mountains, yet I'm not totally happy with it, mostly since there are only the original unreal and shane mountain textures to my knowledge.

[Swanky]

Final version 1.0 release!

This will be my official release for the contest; download link can be found in the first post.

Changelog from Beta 1.1
I added in Revelation's footstep sound manager, had some more sorcery done on the mountains (upper edge of the mountains now features a depth of field like effect) and optimized the pathing network a bit more. Bots are now making better decisions whether to jump off a higher ledge or not, leading to an overall better spread out bot density across the map (they won't jump down towards the Pulse, Shock or Bio guns as often as before but also check how much health they have, whether they really need that armor and whatnot). A few more performance fixes have been made as well such as deleting non needed brushes and selections outside the playable area.
The Skaarj panels have been changed to something more fitting.

[Revelation]

There's a slight issue with my FootsteepSoundManager here: For some reason, the surface tracing doesn't work properly when walking on a lift, resulting in the Default-Sound being played. Don't know why this happens, but it's easy to fix: Export one of the Footstep-Sounds (I'd use Metal1_F) and reimport it as MyLevel/Special/psound1, thereby overwriting the original Default-Sound.

[Swanky]

Shouldn't be a big issue since the liftsounds themselves will also play sounds at the same time. You really wouldn't notice unless you take the time to listen to it. Won't happen in a regular game. Not sure I will / can fix it until tomorrow. Thanks for letting me know though!

[Revelation]

Still sounds pretty weird to me... I've quickly made that change and sent you a hotfixed version via Blizz-app, would rather suggest to replace it in the zip-file.