.Map :: MoHAA Mapping

Ok im sure its been asked before but cant find it in initial searches.
The map that i've virtualy finished is nearing the time for a full compile and im contemplating the usefulness of full vis compared to fast vis.
A while back I did a full compile and it took around 2 days (cant remember what the vis count was then) and it seemed minimal if any performance improvement over fast vis.
The stage im at now the vis count is 600,000 which isnt a problem..fast vis takes around 4 1/2 hours in comparison to I estimate 4-5 days full compile.
Bear in mind please that I have correctly used vis leafs and area portals and after fast vis the fps are great at least on my system (not under 100 fps).
Sooo after all this do you realy think its worthwile to go through the laborious process of a full compile of which take up so much time?
Any feedback from the optomizing king (Bjarne) and other expierienced mappers would be greatly apprieciated.
PS heres a few new screens.




Hope you like

Cant offer advice on vis times but it looks the nutz m8.... cant wait to test it .




Llama aka Seere aka Raum

ahh hello m8,thx for the comment..i need to contact you at some point.
If I dont see you on the server ill get yer msn over the weekend.
(sorry for private chat)
Seere.

Can't really answer your questions but MAN those pics look GREAT
Nice work m8. That first pic looks like a photo.

According to some there is no difference between fast and verbose. I have noticed that verbose bsp, vis, and light will make the resulting bsp file slightly smaller in size. As you probably know, the vis compiler is not very smart to begin with, so fast vis can produce more vis problems. But, if your map looks and runs good after a fast vis compile, go with it.

Bah! I think you mean fast and a regular compile. There is no correlation between fast and verbose, verbose simply outputs more data, hence the name. Using verbose should not effect the resulting file. If the the FPS is that good, I think you are fine with a fast compile, just do maybe a -final and -staticshadows on your light compile...

ok gents thanx for your replies,apprieciated.
Seere.

Looking sweet, seere.
If there are no leaks and the map is cleverly designed with Vis (and lines of sight) in mind then maybe a -fastvis can be sufficient. Sometimes a map may look simple in terms of Vis to the human eye however it can be complex to the compiler. For instance if you've made a skybox that stretches out over the whole map there will be lots of extra leaf_portals. The compiler calculates which portals can 'see' each other and which ones can't. The more brushes/entities(models,doors,windows), the more leaf_portals which in turn create greater compile times.
I've never tested visgroups with -fastvis cause I didn't think they would work properly. You probably know this but you can test it by using the 'r_showtris 2' command to see what's being drawn. But if you're sure they work and the fps is okay then go with -fast. On the other hand if you have a very powerful pc and vidcard you may not experience slowdowns on the map but others may(?). If I were 100% sure my map is finished I'd wait it out and go with -fullvis; but yeah, 2 days seems a bit of too much. I'd probably save time on Mohlight and use -fast lighting since I've never seen any difference with -final, except for ridiculous compile times.

if you are 100percent sure the map is done i would do a full vis compile. theres no reason not to i mean u can still use ur comp and it ensures no vis problems. although if u plan to release it just for fun for a clan or something i would do a fast one since ppl who play online usually have pretty good computers. its all up to you.

Thx once again ..yes i do have a pretty good stystem/graphics card however i benchmark my work(fps) on what i get on the stock maps...its comparible so I'm happy from that perspective.
As its a very similar layout to southern france its a great example of how a map should be laid out...ie structual brushes/vis =easy to implement.
In response to the other reply i'm kind of hoping it will be more succesful that a "clan only map" but then the final verdict on that will be when I release it to the general public.
I guess having put so much time and effort into the map I may as well go for the full compile .
Hope some of you guys try this when released....*scared but quietly optomistic that you will enjoy*
Thx again.
Seere.

well the screens got me interested. ill giv eit a dl.

seere wrote:A while back I did a full compile and it took around 2 days (cant remember what the vis count was then) and it seemed minimal if any performance improvement over fast vis.
The stage im at now the vis count is 600,000 which isnt a problem..fast vis takes around 4 1/2 hours in comparison to I estimate 4-5 days full compile.

2 days!!! ouch.!! wats your comp like.
i done a full compile before meself and the longest it ever took was 3-4hrs. and it was on a fairly big map.

but any way if you want to lower the vis count a bit try detailing a few brushes. but make sure you detail brushes that are details eg. tables, benches and the like..

Yes I appreciate making things detail lowers the vis count however i'm not concerned with the size of vis at the moment as its just over a quarter of whats allowed(2.2 mb I believe)and anything that should be detail is.
The main question I was asking was the benefits of full vis over fast vis ,in my testing there isnt much.
Anyways my questions have been answered and thank you all for your input.
Seere.

I think a fast vis leaves the player seeing more. I found better fps when i did full compile on my levels.

Fast vis is the devil.



First: Are you sure you know how to use detail/structural brushes correctly? If not: Read this tutorial on effective detail/structural design made easy

If that part is made correctly, it COULD be a good idea to delaet all vis_leafgroup brushes ( ).
Because man vis does slow down compile a lot, and is often used incorrectly anyway

If you feel, none of this works ( or it is already done correctly ), mail me the map file and I'll have a quick look at it.