As some of you may remember, SlappyThePirate create a custom edition of the halo 3 valhalla, featured on this site, just before he left. It was a great map, good detail, decent tag set, nice shaders, ect ect. Altogether it was an awesome project, and a great job by slappy. But the map had some issues that were never resolved.

Biped

Hopefully several permutations, improved shaders so that they look sharper and better multipurpose, reflection and detail maps so that they look more like they did in H3(WIP)
Animations(WIP)
Ragdoll(concept WIP)

Vehicles

Boarding and hijacking(Easy enough, but still a WIP)
Machinegun turret(replaced firing sound, fixed firing so it starts out slower and speeds up to maximum, hoping to make it detachable)
Missilepod(completed concept of third person missilepod, polishing and then will use the same method for turret)
The fusion coils sunk through the ground when you touched them(fixed http://www.xfire.com/video/5252bd )
The wraith was unnecessarily high poly(fixed)
The wraith's turret wasn't attached, it just floated in the air(fixed)
The wraith's secondary turret was unusable(Fixed)
The mongoose was almost impossible to control(fixed)
It was suggested that the vehicle needed boost. (Using animations and functions, i think this one is plausible. WIP)
Antenna and crates were scenery, not vehicles, thus making them immovable(WIP)
Banshees camera marker was too close to the vehicle, making it hard to steer(fixed)

Scenery

Man cannons were replaced with teleporters(concept to make man cannons sync is sound, incorporation is WIP)
Trees weren't rigged right, thus making them look like they were breathing, not swaying(WIP)
Waterfall was a flat plain with moving texture and particle(replacing with a rigged piece of scenery so the mesh can move and using shader that fades to misty effect, then particles below that to give the effect that the water is turning to droplets as it falls)
Base flares(WIP)
Blue base flag stand was red(fixed)

BSP

BSP was wasting pollies as well(WIP)
Phantom BSP(fixed)
Bases had generic metal texture for the interior(fixed)
BSP shaders were a little too cartoony (WIP)
Water was too unrealistic and used to many shaders(fixed)
The decal on the glass for the basses was a piece of scenery(replacing with UV'd glass texture with bump map of the same image(WIP)
The BSP had 12 different shaders using 12 different bitmaps that were all identical(reduced to one)
Hidden room(being replaced with a credits\ester egg room that will also showcase a custom gametype)

Sky

Sky was too dim of a blue (WIP)
Sky needs mountains in the background that fade a bit so it looks like the level continues outside of where you can see(WIP)
Sun lens flares didn't suit me(WIP)
Skybox was set up so that you'd just run into a wall(set maximum height in BSP so that you can only go so high, but you don't hit a wall)

Misc

The backpack weapons would stick through the warthog seat and banshee when you entered them(fixed)
FP legs(completed concept, Polishing WIP)
BR was stupidly high poly, wasting processing(fixed)
Gametypes(should be all gametype friendly)
Custom gametypes will also be incorporated so that we can have something to mix up multiplayer(WIP)

I think that's just about it. I've been working on fixing and patching these issues, to get the map to it's full potential. I'm making this thread because the community seems slow lately, and i want to liven things up a bit. This thread is to showcase the things I've completed, I will update things as they come along, and make sure to give credit to anyone who helps. I don't have any pics or anything to show atm, but i'll try to get that all up soon while keeping progress going. Let me know what you think and we'll see if we can complete this project soon and have a decent map to play while others work on getting some more updatage together to show.

Credit goes to slappy for the original project. I'm not trying to take credit for this, the only thing i'm taking credit for is the work i'm putting into polishing it into a finished piece.
Edited by grunt_eater on Sep 27, 2013 at 11:25 AM

sounds like quite the project. good luck on the remake. though how are you possibly going to do ragdolls? not to mention fp legs might lag about quite badly depending on someones connection.

Fakebumping may be an option for the bipeds.

If you do decide to do that, be subtle with it. Make sure it's not obvious that it was fakebumped.
Edited by Dumb AI on Sep 26, 2013 at 10:35 AM

The ragdolls will be fake ragdolls. Using ragdoll simulator in max i create animations, the animations are then used for death animations in halo. All in all it should look okay.

I'm looking into doing a full reanimation. Not just because i'm already doing the death animations, the flamethrower animations(for the missilepod), the aiming animations(so they don't look ridiculous). Ect ect. But because i think halo has been in need of something like this for quite some time, and i think it will add a little flare to the map that it needs.

Yeah, basically what i'm going to do for the bipeds is build it in layers. First sharpen the base map(maybe fake bump) Get the color values how i want them there, next do the multi, get the right balance of detail\reflection performance there, and alpha channel depth so the colors look good(so they look not as cortoony). Next use a custom reflection map that can give the effect of specular(also thinking about maybe having two shaders overlapping, first with all of the other bitmaps and a simple reflection and second with no base shader, a more specular look, and fake bumped so that it looks good.) So yeah, we'll wait and see where it leads us to.

Also, due to my increasingly busy life, i don't have as much time to work on this project. So i may post every now and again asking if someone could help and give them credit. And keep an eye at the first post, i keep remembering things that i need to add to it.
-Boarding and hijacking(Easy enough, but still a WIP)
Edited by grunt_eater on Sep 26, 2013 at 11:30 AM

You don't want the legs to sync, you want the clients to worry about themselves. Also, having a MP map, you WILL have visual errors with the legs without host testing; people will see the legs on the bipeds, so unless you have the FP legs as the only legs on the biped, there is gonna be visual problems (Normally, I'd say who cares about the visuals, but the entire point of FP legs is visuals; it adds no new functionality)

Would you need to worry about other players seeing the PF legs if they had a second, invisible, permutation set to switch to when you were like a foot away? In this case, they don't actually ever get close enough to switch to the visible permutation, but the player with them attached is always that close. What do you think?

That should work (I think).

Maybe you could do that via lods

It would reduce the cases of it, but people will still be able to get close enough to see the visual errors. These same visual errors will cause problem if someone is camo'd; they get close enough, then suddenly a solid pair of legs showing up on screen. (and as the other guy said, it would be a LoD; if it was a permutation, you'd need to know who the local player was anyways)


Anyways, you gonna bother showing the shields on the FP legs? If you are, you'd want to make the leg parts as bipeds (purely because of the number of script related things you can do to bipeds compared to scenery), and since your not using OS, check for shield changes on the player, and then use the flashlight state on the leg chunk bipeds and instead just put the shield function block as the integrated light state instead. For camo, without OS you'd either replace the legs with an already camo'd variant, or just make the legs not show up during camo. Anyways, for a non-OS user, you have a lot more to think about.

the main thing that wore slappy out was the man-cannons, he spent so much effort trying to implement a way for them to sync online and work consistently. he had a good idea but he just couldnt get it to sync.
it seems that no-one has made a man-cannon that works consistently, on one of the earliest versions of valhalla the man-cannon would dribble u out in front of the base ur trying to escape from or send u accross the map, smashing u into the side of a cliff.
id say leave the teleports unless kirbs has got a trick up his sleeve.

i wouldnt worry about fp legs, sounds like too much effort for too little gain.

also, jacking gonna sync? i remember cmt saying that without sauce there were issues.

Edited by Ubergoober on Sep 26, 2013 at 06:37 PM

When i said permutation, i actually meant LoD. i just didn't remember the name. As for making the pieces bipeds, i like that idea. I wonder if there would be a way to do it all with one set of legs. So that everyone sees the legs that were created on their PC's on themselves. i mean, with local player detection, i guess. But i'm still not sure of an efficient way of doing that without screwing with gameplay.

Ubergoober: I believe kirby is the one who gave me the idea of this method in the first place. It's a vehicle with an entering animation of the fly-through. You enter a trigger volume, it loads you, then waits for the animation to go all the way through and drops you off at the bottom. While in the cannon you can actually still shoot as well.

As for boarding and hijacking, i don't see why not. The way it works is with a vehicle as well. Say you want to board the wraith. There is a "boarding" vehicle attached to it that you can enter. everyone's running the script that attaches the vehicles, everyone sees you in the vehicle and everyone sees the vehicle moving where the wraith is. Would it not sync?
Edited by grunt_eater on Sep 26, 2013 at 06:48 PM

It doesn't matter if everyone's running the script, the object names do not sync, so unless you create client side doubles, the clients copy of the script is doing nothing. And you dont want client copies littering the map for that sort of thing, as it wont serve your purpose in the slighest (Where you do want client doubles, is in the mancannon script, for reasons ill explain shortly). So since the people are entering the hosts vehicles that the host is calculating location for, they will follow the vehicle, however the game will only update their location every half a second at best. And, if you're forcing people into seats after the boarding, that will also be a problem; Vehicle names aren't synced, only the host will be putting people in the vehicles, and that itself does not sync; the host never informs the players that it has placed them in a vehicle. Now, if you had a client side copy, the client could put them in that client side copy, but that isnt the copy your worried about (This however, is used for the mancannon situation. Since it doesnt matter whether they're in the real one or a client copy, as they'll always be in the same place, and entering does the same action regardless of what PC it is, etc)

Sure, hoping onto a side vehicle and killing the driver will work, but it will look bad, and if you put them into the driver seat afterwards, they'll be unaware of it.

Oh. I see. i had no idea that object names didn't sync.

And that just broke the bubble..

Yeah, what happened to Slappy? Haven't seen him since 2011's or so.

That's nice.

First fix i made was the fusion coil physics.

http://www.xfire.com/video/5252bd

Hey, kirbs. If you have a piece of scenery that's playing an animation, and it runs the (object_create_anew <object-name>) command every time a player joins will the animation sync?

And what if you create the boarding vehicle off screen, then attached them? would the host see doubles because of clients? and would the clients see doubles because of the host?
Edited by grunt_eater on Sep 27, 2013 at 12:16 PM

The host never sees doubles, only the client sees doubles. And the clients running the scripts, are only going to effect the doubles, never the real ones. So, they'd have their boarding vehicle attaching to the side of their copy vehicles. If the script ever gets called on their side to load into a vehicle, they'll get loaded into those doubles.

The only way to have a client call on the host vehicle is OS and storing the object in a global variable with (player_data_get_object 0 nearest_object) when the player is beside something.



And for syncing your animation, I myself would just resync the animation at the end of every loop by having the host kill a biped; when the biped dies, everyone recreates their object and plays the animation again, but sure, what you want to do should work, although I dont think people will like the sudden jolt everytime it's recreated mid-animation.

I think i like that method better. Okay, so the client will see doubles because they're seeing the hosts and the their own? I thought you said they wouldn't see one, which would result in them being loaded into it without their side updating.
Edited by grunt_eater on Sep 27, 2013 at 12:51 PM

The scripts only effect the doubles. The doubles are unassociated with the originals. So, they will go up to the hosts vehicle (the real vehicle), attempt to hijack it with the hosts boarding vehicle, which will be lagging along behind (updating at best, every half a second). Meanwhile, the double is laying somewhere in the map, where the client copy of the script is moving the client copy of hte boarding vehicle too at real time. Since the script on the client never sees them in that boarding vehicle since its monitoring the double rather than the real one, the script never activates to load htem into the vehicle. So, only the host sees the players in the vehicles that the script loads them into, as this only occurs on the hosts side.

But i meant, hijacking aside. Say you wanted to do something like the wraith turret. You have the host who creates it, in a secluded area of the map, and runs it for themselves. Then you have the client who creates their own, same place, and it's attached to the same place. If each person has their own vehicle, attached to the same place, with the same name, and a script just telling the vehicle to fallow the wraith around. Would that work? i mean, the host would see them in the hosts vehicle, and the clients would see themselves in their own vehicle, but they'd still be in identical vehicles in identical places, right?

Nope. When you say it'll attach to the same place, thats completely wrong. The host will attach his to the real vehicle, the client will attach his to the fake vehicle. Sure, after the script loads people into both, its going to freak completely out, but it will start moving them into the general area of each other assuming you don't do what you said about attaching, but it will be visually horrible, and you cannot have physics on that turret, or else its going to ram the real turret and fly away; in the case you actually have the client double moving with the vehicle, it will still only update at best every half a second. It would be fine if you have it on a practically immobile vehicle, or a completely immobile vehicle. In those cases, you can have the default permutation nulled, and the hosts and clients set the permutation to the visible on their own copies, although that still causes minor issues for when the client isnt in the double, but the actual hosts (aiming animations and such), but it would be less visual issues than having two stationary turrets in the same place.

I see. I just don't understand why if everyone's running the same map with the object names compiled in it doesn't transfer.

Because the engine was never setup to transfer it. There was no reason what so ever for Bungie, gearbox, or anyone to send that data between PCs. No default MP maps had scripts, so having a script system that would sync or effect all objects equally was never part of their intentions, they only needed any scripting for SP, and even then, it was only for the local machine.

What if you attached the host vehicle and the client vehicle to the player in the wraith? Then you wouldn't need the wraith to have a name, just the object, client creates a now one when they join and script waits to attach it to a marker on whatever biped gets in the vehicle?

PS: Working on the BR. Reduced poly count(no longer ridiculously high), and mimicking shaders to a direct screen shot from halo 3.

Except the client cant see when someones in the wraith, and you can't just (vehicle_driver <vehicle>) on the client side, since that would be looking at the double rather than the real one due to the name not syncing. If you're moving everything via player, then everything your moving, will need to store at least 16 possible values via killing bipeds for EVERY seat that triggers this, to tell the clients what player to move it too. So, even 4 vehicles with one seat each, that's at least 64 values (if you plan to have no killed bipeds as your forcibly exit, no attach/detach movement) that you have to hope are accurately depicted to the clients. Unless you know binary (or, use the OS command that deals with binary logic for you), your gonna have too much trouble (Although I do have my device sync script out there; read it, it will help you understand how to store the value as binary in bipeds)

i wanted to wait until i finished it to give this update, but i guess this is okay to show.

Old


Mine


Halo: 3


Still to do, finish modifying diffuse map for the body of the gun. Edit shaders, reflection ect. And the problem i'm having, which is why i'm posting it early. In 3dsmax, i have two sets of numbers plates. One for counting in ten's, the other in one's. The one for the tens isn't exporting correctly. I mean, i looked in the JMS file and there are two of them there, but when i compile it's only one. This causes problems for the ammo counter. The reason i'm posting this early is because i need someone's help in order to complete it.

PS: my BR scope has the simple sun reflection and a full panoramic of highground. The sun reflection actually matches up with the sun positioning in the map.

PPS: after i posted that picture i fixed the bolt's smoothing.

Maybe they are using the same shader and they get compiled as one (not sure if that would actually happen, I got little to no experience with materials)

The numbers on your BR are aqua instead of blue...maybe you can fix that later.

Get a full panoramic of Valhalla :P
Makes more sense that way.

Edited by Dumb AI on Sep 30, 2013 at 09:17 AM

On a less technical side, the Battlerifle needs to be closer to the screen and the reflections on the surface need to be beefed up a bit to show that lights actually bouncing off it. The counter needs to be cobalt/light blue in order to match the halo 3 one. The FP arms have little reflection variation as well. I'm assuming you're using vanilla Halo to showcase the new weapon and therefore, the vanilla halo 3 spartan model? Maybe in the final build, you could use those super amazing open sauce Spartans that rivaled the halo 3 ones that Were lying around the forums for like a month or so. I'm sure you could find them if you click back a few pages.. Also, please make the map well lit.. So many times with these Valhalla maps, the sun has like no influence on the lighting of the place.. Like, the textures will be dark in the areas it should be light and then the shadows are 'too' intense in the areas that should be dark.. Then again, I always played Halo with high Gamma..

did he play any better compared to when he was straight?

great improvement on the original he who eats grunts, keep up the good work :)

Why are you even using that model? It's so old, and it's the Halo 3 beta model, THIRD PERSON. Look at the scope. See the red tipped knobs? No, because they're not on the TP model. It is low res compared to the FP model.

Not on my watch. Unlocked.


And I agree @Mootjuh,
Edited by AllySuzumiya on Sep 30, 2013 at 06:33 PM

I wont jack the thread. Don't worry about it.

@AllySuzumiya: First, i'm not adjusting the animations yet, so the position isn't going to be changed for a while. Second, i'm working on defuse maps, at the moment. Reflection and multies will be done later. Third, yes, i noticed that as well on the side by side comparison. I will fix it soon. Fourth, this picture was posted to show progress on the weapon itself, not the arms. Fifth, i'm not going to be making an OS version of this. I myself can't run the graphics upgrades with it, and i don't want to split the player audience by making it so that only OS users can play it. So i don't think so. i will make sure to get a good feedback from the community before i make anything final.

@Jesse: I'm using that model because it's the model i had. I had to reduce poly count because it was stupid high. If you have a better one then i'd be happy to check it out.

@Mootjuh: I'd actually rather make my own. And sure, i'll go back and mark what's done and in progress.
Edited by grunt_eater on Oct 1, 2013 at 01:59 AM

You wouldn't be splitting the the player audience by remaking a map 10 people else already remade. I could link you to like 5 Valhalla maps off the back of my head. They're as common and sandtrap mods honestly. Also, OS is nowhere near as demanding as you think. When my laptop worked, I ran open sauce in full fidelity with upwards of 30-45fps. My computer is 7 years old. Granted, it also runs Black Ops 1 but that's beside the point. Try getting more RAM

And i can completely guarantee you that every one of those valhalla variants will be absolutely incomplete. I've seen them too, they all have problems, they're all pretty much crap. That's the entire point in me making this variant. A version that has completed everything within it. Not 12 concepts crammed in a map with halfassed tags and an ugly BSP with terrible shaders. I'm making this so that i can say that there is a good version of valhalla out there, one that looks like it's had a lot of work put into making sure that everything is good, well done, and aesthetics are as close to halo 3 as this engine can get. And i don't mean i'm trying to make halo 3 in CE, because i know the limitations of the engine. My concern is making this map look good.

As for OS. I have tried it, the version before last just made halo exception, thefireyscythe tried to help me by sending me d3d9.dll's or whatever and each one disabled another component, and when we got to the end and i still had the problem, i gave up. Now there's a version that runs, and all of the visual upgrades are disabled. So what's the point? Anyway, i know just how demanding OS is, i know because I've tried. It's just not going to work, and not because of ram, because my video card isn't compatible. I'm planning on getting a new laptop this year or next, but this map will be non-OS. That's what i want for it, because i'm not going to force people to decide weather to get OS just to play the map, or to not play the map because they don't want OS, and screw over the people that can't use it. [/rant]