Question about the build concepts behind Goldrush

(Ifurita) #1

Anyone looking at the backside (void-facing) of Goldrush will notice that is is pretty complex. I imagine most people here would have made a big hollow box and then built all of the GR architecture inside. Why was GR built this way and what are the advantages?

I was discussing this with Hummer last night and I suppose that:

  1. Less volume used in the map - not sure if this really makes a difference
  2. Potentially reduces the PVS thus reducing the vis table and compile times, but since everything is sealed, would this really matter?
  3. Potentially eliminates some brusheswork and tris, but again, since everything is sealed, would this really matter.

Just curious, because it seems like it took a lot of work to make sure you had everything sealed properly

(hummer) #2

I think, mainly because big boxes = lots of extra portals and space. That means bigger BSP sizes and longer load times. Larger BSPs = longer downloads, and longer wait times = unhappy gamers :slight_smile: Damn, I need to lay off the = sign.

Anyway, that was my guess :slight_smile:

(Ifurita) #3

yeah, could be. Especially the vertical portals. In any case, I wanted SD’s take on it since it’s sooooooo much different from the simple box that I have :frowning:

(MindLink) #4

I guess goldrush has such a complex design because SD (unlike many new mappers) used a “tunnel”-approach to designing the map. They didn’t start off with a huge box and added buildings as need but rather imagined each street and passage as a closed “tunnel” like in the good ol’ Q3 days, just that the ceiling shader is a skyshader instead of a normal one. This is much more complex to build but pays off in terms of speed.
Just think of the streets as rooms connected by passageways and you’ve got a good idea of what I’m talking about. :slight_smile:

(Ifurita) #5

No, it’s more complex than that. Open up Goldrush and look at it from a void-side perspective. What you describe makes perfect sense but if I were doing this, it would look very much like 5-6 larger skyboxes connected by tunnels

(MindLink) #6

It looks like the mapper started with a floor and just added parts around until the section was closed from the void, not taking to much care whether it’s nice from the outside or not, but since it connects to the void all those edges and complex structures outside are completely irrelevant anyways.
You don’t have to build a room or a passage starting with a box, you can do it by simply mapping parts until you’ve got a complete closed structure. This is indeed more complex and will result in many errors for the unexperienced, but it IS a possibility to do it. And when you do it like that you simply don’t care how the outside structure is, you just try to keep it closed, best without adding any additional brushes.
Like this you don’t have to worry about filling all unused spaces with caulk and stuff (so vis doesn’t put any portals there).

(Ifurita) #7

Yes, I can buy that and that accounts for some of the layout. But there are other areas where a single solid structural brush (or what could have been a single brush) looks like it was hollowed out and the piece of skywall where it meets the brush was deleted

(sock) #8

GR went through alot of changes and revisions. (The final file on my machine is version 268). It was designed intially from a Max model which was scaled up until a player model looked right in the map. (The max model idea was an idea to see if concept maps could be created quicker) Once the model was ok I converted the model into basic brushwork and set the intial detail style by creating the bank area. After this is was worked on by four other mappers who did varous parts of the map.

At one point the map reached over 20Mb in size and I spend 2 weeks culling and optimizing alot of the street designs. Its sort of like clipping a map, its damn hard work and not really much to show for it afterwards except alot smaller filesize and maybe an easier map to work with. I also converted the temporary floor brushes (mirroring the max model) into a terrain mesh so the sand would blend and feel more wind swept. There was no height map or easygen trick, it was all done by hand/eye and imho produces better results.

Every mappers who worked on the map left their mark and its probably why it looks so nice. When a map is shaped by several people it gains a certain lived-in look and feel. All the brushwork was changed and adapted as each mapper worked on their section. All seams had to be blended together and the map was given a final detail pass. There is alot of history to GR and it was one of the intial maps to be turned into a concept. I’ve certainly got some very fond memories of various sections of that map.

Sock
:moo:

(Ifurita) #9

interesting. I, as I am sure everyone else does, like to understand the thought process behind someone’s creation, since everyone approaches the project in a slightly different way.

I’m sure the ‘start from a box’ is fairly inefficient, because it’s unlikely that you really know what size map you want at the start of a project and humans have a tendancy to want to fill up available space with something

Thanks for the explanation

(Loffy) #10

Thanks for telling, sock.
Goldrush is such a great map.
Obviously there was alot of hours of work.
When I “fly into” some of the buildings, in radiant, I see fireplaces. It is as if the house/floor was built initially to be open. But then in final release the building remained “locked”. So the only way to see that fireplace (all made of caulk) is to fly in there, in the editor.
// L.

(CptnTriscuit) #11

Did you find that building the concept map in Max was worthwhile, or in the long run is it better to just stick with Radiant? Or somthing else mabie?

Concerning final maps looking like the ‘big simple box’ VS ‘tunnels’, my maps usually come out to be some weird mixture of both. Usually it starts out as one big box, but that gets changed and added too and redone constantly as I move from one part of the map to the next…to be honest I never gave it much thought. I guess its a balancing act between using map space efficiantly and good construction techniques. …or am I way off? :smiley:

(G0-Gerbil) #12

So go on then, dish the dirt - what were the easy offiical maps to build and what were the simple ones? :slight_smile:

How much planning did you guys do before you even started the map?
And how much of the initial planning tended to end up in the final maps, after playtesting tweaks?

I’m curious what ‘rules’ you use to decide on a map layout, maybe something like:

  • Each point of interest must have (2 / 3?) routes to it
  • Routes must take roughly the same amount of time

etc?

(fattakin) #13

Nice explination sock - friggin good work so glad you are happy looking back!

(Loffy) #14

On the issue of looking back. Here are some screenshots I found on a Swedish wolf site.
http://www.swertcw.com/default.php?c=media&p=et
You can see the alpha versions of some of the later released ET maps.
// Loffy

(bsimser) #15

I always find it more interesting to see the evolution of a map/game/idea than the final product.

Any chance of SD releasing alpha/beta versions of ET with these original maps in them? Might be a hoot to see how things might have been or works in progress and how they morphed into what we have today.

(Blackadder_NZ) #16

How about releasing this map for the mapping community? Or give a quick speech on some maps which got far, but in the end were scrapped?

(flashkillaman) #17

Did they want to make et like this? I think then it wouldnt be such an succes

(Blackadder_NZ) #18

I’m just interested in seeing what maps SD made for ET that were scrapped. There must have been a few…?

(flashkillaman) #19

some more pictures i could find

(sock) #20

The map was from the SP element of the game which was developed by MadDoc not SD.

There was 3 additional maps which were dropped because of time limits and resource issues. Unfortunately I cannot show any screenshots of these maps and they will have to be left as a mystery.

Sock
:moo: