CMP to SUR Conversion Tests
-
StarTrader wrote:
I have rebuilt one of the models completely a couple of times but still get asymmetric surs for it - the ship has forward-swept wings, the left wing’s sur connects the left wing forward tip directly to the nose of the ship. The right wing’s sur is correct.
So the ship looks like the Phantom of the Opera or Dracula from above, with its left wing holding a cape up to its nose!
I will try to define that model’s problems more accurately for you and will post it up if I can find what in the model is causing the asymmetry in sur builder.
OK? Couple of days.
-
ST, you old bag, quit harassing the guy. I didn’t just throw in the towel, I threw in the kitchen sink, the laundry basket and detergent to clean them
Actually I have been quite surprised at how well he has continued this project, if you saw the code you would know what I meant.
Ok, since you have mentioned it, I have been contemplating on releasing source code for every project that is FL related I’ve made, would that make ya happy?
Anyways messing with ya Star Trader, I just couldn’t resist, btw, I still have that hedge trimmer for your beard just waiting on ya…
BTW, one reason Ive never released my code is that in the wrong hands, it can be very dangerous. There is code that could easily wipe your hard drives in a matter of seconds, send all of you personal info to any ftp site, rewrite entire folder structures, wipe your registry and even more malicious stuff. This is why I only want to release it to those I know who wouldn’t use it in that manner. If Bullwinkle wanted to he could write a short routine to put swirlies in Star Traders beard JK…
Compentent programmers plz send me a pm on the project you would like to take over.
-
Well if it’s just that I would be interested in the code snippet that creates the tree of the Bitsection, so that I can compare it with my approach. Would that be possible?
-
StarTrader wrote:
Couple of days.OK, Cool.
The goal of the SUR Builder is to quickly create decent SURs for most of your models. If you have an oddball model that requires a hand-crafted SUR, then that is outside the scope of the Builder. Why hold up development of your mod just for one model?
All other reports are either that the Builder works as expected, or the model was bad to begin with.
My understanding is that there are at least three other workable techniques:
- Resize vanilla SUR
- SUR-splice
- Hand-made SUR in a modeling program
If there is more to it, then teach me.
-
Oi!
Hold on a mo LS, stop running away - the sur builder is connecting vertices of two entirely separate cmp groups!!
That’s what is causing this problem on several of my models, not only one or two, and large and small ships alike. I am trying to investigate if it can be prevented by regrouping the model.
I have focussed on one model and completely regrouped the ship twice, and it still fails, producing the “Phantom of the Opera” sur! The left wing and nose are in separate cmp groups and sur builder is still connecting them. It should not do so, am I right?
I will post the model once I know it is not a model problem.
-
I ain’t harassing, just want to get the builder to be better. As I said I appreciate BW’s work, I always do of anyone who puts so much effort into a spare-time no-earner project.
-
No I don’t want your source code, but others who want to develop new releases or sur utilities or exporters will, or they will be starting from scratch every time, am I right?
And you will agree that the sur builder is the most problematic to create, so far? No-one has helped anyone else in this until now, hence we have 2 sur export plugins that don’t work reliably, and sur builder may be easily improvable once we define the problem better…
At least I am trying, and it’s not only for my own benefit.
I could give up, but it’s in everyone’s interest if I can find the cause - if it’s a group naming problem, then it could be an easy fix in the builder. If it’s the way the groups are structured, it will be good for everyone to know so they can fix their models.
So don’t get on my case so darned fast! Walking away is easy, don’t tempt me more.
Eeee-bloody-yaw!
-
-
Hey I like giving you a hard time, Im from the volunteer state and volunteered
Anyways, I haven’t even looked at the code in over 2 months so I couldn’t even begin to know whats causing it.
One possiblility is if it isn’t reading the mesh properly out of the model, don’t know, it never happened to me.
BTW, I have to keep your wits sharp, can’t have you getting dull on me
-
It’s OK, no probs.
Just hang around a bit to help if needed bud.
I hope to find a way to regroup or rename the cmp groups that helps sur builder to make a good sur for the same ship, If I’m not successful then I’ll pass the model to BW, so you guys can try to find it. I’ve done the obvious - fuselage, left wing, right wing, fin - but as I said the sur is the same, bridging the left wing to the fuselage nose.
I’ll try to get the diagnosis completed asap.
Thanks.
-
Hi guys, here’s the diagnosis and findings…
I do have a problem with the sur built by the sur builder for the Goran, my test ship that had the “Phantom of the Opera” look.
It’s not making all of the sur parts for all of the Goran’s cmp groups, just as with Bbalazs’ terrain model.
The initial problem that I had was most likely due to some unwelded vertices and open cmp groups that I found on the ship model.
That is not happening now, I fixed those by removing all unnecessary model parts and regrouping until I have only 6 distinct and simple, welded cmp groups:- hull, left wing, left wing tip, right wing, right wing tip, and fin.
Then I ran sur builder and there are two problems in the generated sur file (I tried altering the Sort selections and Duplicate radius values and remaking the sur but the problem remains)…
1. The “right wing tip” sur part is missing entirely.
2. On the fin sur part, the front vertex at the base of the fin is missed, so the sur sweeps back to the next 2 vertces and the front part of the base of the “real” fin is exposed.I attach a zip file with the ship model .cmp, the generated sur, and the ms3d model with both the ship groups and the imported sur parts using adoxa’s sur importer, and the sur importer log too.
Look closely at the fin with the fin sur part selected and you will see the problem with that.
Hope this helps to find the cause of these problems.
Thanks guys.
-
Does it work in-game?
-
ST even if they were created they wouldn’t register properly in FL, in my experience single vertice edges like your wing tips always act like a knife and cuts through every other sur, and would probably only be stopped from going clean through when the main body hit the other sur.
You need to split the “knife edge” on the sur model and move the top and bottom vertices apart, then fill in the holes with a couple of polys.
I haven’t used the sur builder and I probably won’t as I no longer mod, but it wouldn’t surprise me if it doesn’t like “knife edges” either.
-
I tried a bit on your model. I tweaked a bit on your model and got the SUR-Builder to do a proper sur.
I rebuild the model a bit on the front bottom cause it had there some strange poly´s, same on the cockpit. On the Fin i closed the object. On the tip´s of the wings i tessalate the poly´s to have a few more vertices and faces.
I attached for you a zip containing the .cmp, .sur, .ms3d so you can take a look for yourself.
Hope this helped.
cheers
CURSOR -
Thank you very much guys, appreciated.
BW - its hard to tell on a fighter like this one. The problem will come on large ships where the missed part could be the size of a base.
So it is important to find out what is causing this problem - remember Bbalazs had it too.
If it is the knife edge then of course we can avoid this in a model, or if it can be done by the builder, to set a minimum sur thickness of 0.1 unit (10 cm) at any single-vertex point, then that would fix it every time for everyone.
Get LS back on his head one last time
Many thanks again guys.
-
If it works then there is nothing wrong with it.
-
When I make an automatic shed-building machine I won’t be happy if it leaves the door off, even if it is only now and then.
There are bits missing from the sur files, BW.
Not only from this model which is not important, but from BBalazs’ blooming great landscape. So it is a common problem, and the cause needs finding.
As I already said, in a small 5 metre ship it’s no deal.
BUT on a 3,000 metre dreadnought ship a missing hole 200m is a BIG problem.
Give me an ear here, I am not being pedantic just for myself - I know the use this is needed for, it is VERY important to FL modders, but it is not 100% done just yet.
Thanks for understanding, I realise you may not have the knowledge to find it, and like I said before many times I am not criticising, but on the contrary I am very grateful to you for your work.
But please talk to LS or others and try to make this perfect as possible, we have far too many 90% utilities where we need to know the “get-rounds” to avoid many hours lost.
Thanks.
-
I am not ignoring your report, StarTrader, but … I think that you are trying to do something that you should not.
The fact that you are able to make a CMP that fools the SUR Builder does not mean that there is anything wrong with the SUR Builder. It simply means that you should use another tool, such as splicing together some cylinders.
If the SUR works in-game, and is vaguely near the correct size, then it is “good”. A SUR is not supposed to look like the model.
You did not show a picture of your SUR, but it looks very much like the ship, which is above expectations for the SUR Builder:
Remember that a SUR is the computer’s view of the hit and collision area of a ship. In general, that should be a rough ellipsoid:
The fact that the SUR Builder is capable of making close-fitting SURs is great, but SURs should NOT look identical to the CMP. A SUR should look like a rough blob made up of a minimal number of polygons.
-
Ok, i also checked Bbalazs landscape. No bad geometry on first view, Fl-SurBuilder created 5 groups but left out the big plane. So the “error” must have been there. I tessalated a few more faces into it and “voila”, you have a proper sur. See the attached file.
So, yes, the surbuilder isnt perfect. But differently “not perfect”. It is building the sur on a part when the geometry is ok and enough polys are there. The bad thing is it just stops when it encounters a geometry on a part that it cannot handle and dont says nothing about. So you have to dig into it and look at all parts where no sur is build and find the reason behind. Bad thing about this is also that you dont know about the order it does that. So when it stops on a part which it cannot handle, you dont know which part that is and therefor also if the following parts are ok.
Basicly telling here, that it would be cool if the Sur-Builder would do at least a check if the amount of groups in the .cmp matches the amount of groups in the .sur and giveout a warning. Same on the order of building up the sur, if it tells you where it was on building up and it stumble, it makes it just way more easy to identifiy the part and fix it.
just my 2 cents
But again, THANK YOU Lancer & Bullwinkle for such an amazing tool, it just save a a hell of a lot time.
Slight hint : tell them that leaving out parts of the sur is a feature not a bug. M$ does the same.
cheers
CURSOR -
cursor wrote:
it would be cool if the Sur-Builder would … check if the amount of groups in the .cmp matches the amount of groups in the .sur and giveout a warning. Same on the order of building up the sur, if it tells you where it was on building up and it stumble, it makes it just way more easy to identifiy the part and fix it.Thank you for that report, Cursor, and for completing the answer to BBalaz’s question.
(also thank you to Schmackbolzen for giving BBalaz a working SUR)
I will look into the possibility of reporting mismatches in the number of groups. I am not sure whether it will be feasible to flag where a mismatch begins, because I am not sure the program will know that. But I will look into both for a future version.
Thanks!
-
i’m not sure to understand correctly so don’t hesite to correct me please
from now and if we want to use the sur builder, we have to redo a cmp with x groups ? right ?
so to use the sur builder we need a very good knowledge on the cmp format and uses of milkshape, import/export freelancer model, eventually in rebuilding a mat file alsoso clearly it’s not a “noob tool” if i can say, no ?
understand me well, i spent, spend and will spend a loooooooooooot of hours making sur, the most perfect i can in 3dsmax
for who don’t know i’m working on TOW, overfiend and me have done the hitbox, so i know how it’s hard to deal with those devilsso i’m very gratefull for all the work you can do but in my mind the tool is in first, done for all who don’t want to take a loooong time to learn how the sur works
therefore ideally, noob wants to launch the tool, select the ship/base, choose betwen bubble or wrapped and clic “make a sur”
believe me, i really appreciate any initiative to put myself unemployed
thank you to enlighten me