51

Re: A new kind of ball

Yep I use Wings3d, which is really nice.

Mapc's OBJ importer is does not support the entire OBJ spec (the importer code is actually only 80 lines of code), but it's sufficiently general to do whatever you want.  Two limitations: all faces must be triangles, and all vertices must have UVs (texture coordinates).

The first limitation is easy.  Just do a Tessellate Triangles while in Face mode.  I usually save objects in .wings format and do the tessellation only for the OBJ export, since it's easier to model with quads and polys.

The second limitation can be an annoyance.  You have to use either Image Snap mode or the UV Editor to assign a textured material to all surfaces.  Once this is done, you need to rename the material in Wings3d to match the filename of the mtrl file you want.  This is because mapc uses Neverball mtrl files instead of OBJ .mtl files, which have a different feature set.  You can discard the .mtl file that Wings3d produces.

52

Re: A new kind of ball

Krabby Krap wrote:

By the way, specifying a boink file for the ball would be handy.  Sometimes the noise just doesn't match up with the current ball.  I myself am no sound designer by far, but we'll probably need more ogg's, like for Reactor.

This idea will be given serious consideration.  The only problem is that the implementation will need to cut across an abstraction boundary in a less-than-elegant fashion.  I'll think about a clean way of doing it.

53

Re: A new kind of ball

rlk wrote:

Yep I use Wings3d, which is really nice.

Thanks for the info.

If I get a chance before I go away, I reckon I might have a go at making something (right now, I'm feeling a bit pessimistic though).

Cheese
==========
cheesetalks.net

54

Re: A new kind of ball

If no one is opposed, I'd like to merge the ball-test branch into the trunk.  Even if only basic-ball is used in version 1.5 of the game, this code is still a better way of rendering it.  Plus, there are a couple of nice enhancements to the renderer that I'd like to take advantage of elsewhere.  There hasn't been much question of the reliability of this code, so it shouldn't impact anything negatively.

Anybody disagree?

55

Re: A new kind of ball

I believe I haven't tested your latest changes, but I didn't see anything in the log specifically related to fixing the rendering of ball with respect to switches and goals.  Can I assume it's fixed?  In that case just merge away.

56

Re: A new kind of ball

parasti wrote:

I believe I haven't tested your latest changes, but I didn't see anything in the log specifically related to fixing the rendering of ball with respect to switches and goals.  Can I assume it's fixed?  In that case just merge away.

Yes that's fixed.  I'll merge it sometime today.

57

Re: A new kind of ball

Cool.

Mac OS X Xcode project & package maintainer.

If you have some Neverball related files you need hosted somewhere, please send a me forum PM/email.

58 (edited by Cheeseness 2008-01-09 02:09:19)

Re: A new kind of ball

I'm all for it except for the changes in translucency of world geometry when the ball's behind an object (which I'm now not sure if I'm making up or not since nobody else has brought it up). From what I see (or imagine), it looks like that has some impact on the playability of certain levels.

Cheese
==========
cheesetalks.net

59 (edited by rlk 2008-01-09 02:35:05)

Re: A new kind of ball

Cheeseness wrote:

I'm all for it except for the changes in translucency of world geometry when the ball's behind an object (which I'm now not sure if I'm making up or not since nobody else has brought it up). From what I see (or imagine), it looks like that has some impact on the playability of certain levels.

Nobody brought it up, so I mentioned it myself.  My personal stance on the issue is that I'm not in favor of requiring it in levels.  If you can't see the ball, then you certainly can't see where it's going.  Which levels do you think are affected?

Edit: I'm not necessarily advocating the total elimination of the concept.  I could bring it back (to varying degrees subject to added ball complexity).

60

Re: A new kind of ball

I like not seeing the ball through geometry, looks more real to me.
cheese, can you give some examples of what levels you feel are less playable w/out the ghost ball?

No objections to merging here.

61

Re: A new kind of ball

Was this translucency for the old camera view angles? I don't see any real need for the translucency now.

Currently Playing:
Celeste and Electronic Super Joy

62 (edited by parasti 2008-01-09 10:52:59)

Re: A new kind of ball

I've played through a couple of sets with the new ball and frankly the only reason I know the ghost is gone is because rlk mentioned it.  I never even noticed it.

63

Re: A new kind of ball

Specifically, the only one I can think of is one of Krabby's bonus levels (it's sort of a tall maze).
Authors will know their levels better.

One of Mym's ones (the one where you have to run over the switches to get the four pieces of ramp to align - 21 on the Nevertable) looked odd when the ramps were clipping by the camera's near distance, but it could have always been like that and it only looked out of place because I was specifically looking for odd things.

If I'm not mistaken the last of Mym's levels in the 1.4 set has the ball obscured for the first jump as well (though I haven't been following what's in and what's out).

It might seem like a bigger potential issue to me because I pretty much exclusively use manual camera and probably end up looking through stuff more often than anyone using chase camera.

Cheese
==========
cheesetalks.net

64

Re: A new kind of ball

There needs to be some handling for the case where user installs neverball and doesn't get any ball at all, due to having an old-style ball definition in his config file. (Maybe the user installed neverball once years ago, etc)
I think the best way would be to just default to the basic ball if the ball definition in config file doesn't point to a correct, new-style ball...

65

Re: A new kind of ball

This is clearly necessary.

A slightly easier fix would just be to change the neverballrc key from "ball" to "ball_file" or something.  That way the game would just create the correct key on the first run, and there'd be no need to test file formats.

66

Re: A new kind of ball

GRRRH.  I've been spending all evening (and by now, part of the morning...) trying to make a new ball.  No luck. Created a cube in wings3d, exported to .obj and so on. (Trying to make a cube inside the basic ball.) i can get the basic ball showing up, but either neverball crashes, or nothing shows up inside it, depending on what i diddle with in the .obj file.   This shit ain't easy.

sad

67

Re: A new kind of ball

I'm sorry, I have a partial document completed.  It'll be a very thorough document when done.  At the moment I have a couple more things I want to add to the mechanism itself, and I need to finish writing it all up.  As I stated before, the winter break is over and I can't devote the same attention to the game that I've devoted recently.  Please be patient and it will all be made clear.

I'm happy to see someone interested in making new balls though!

68

Re: A new kind of ball

I don't understand what I'm missing from the damn .obj file when exporting either from wings3d or blender. I managed to pare down the reactor rods so there's only one rod left, and when i import that and then export it, and twiddle with the mtrl stuff, it shows up. but when i create a cube, export it to obj and diddle with it, it's no show. Frustrating. And I have no idea why.

(Also, the user interfaces in 3d modelling programs seem to make me froth from the mouth, but I'd better not go there...)

69

Re: A new kind of ball

RLK, I'd settle for a quick "in wings3d, click this, then that, then this, then export, and do s/foo/bar/ in the file and you should have a cube" as I do get the .mapping stuff, which is simple enough that I can edit the ball maps by hand, but the .obj ... Ugh.  :-D

70

Re: A new kind of ball

Just fiddled around with it.  My experience with Wings3D and UV mapping is nearly zero, but I did manage to get a half-textured cube to show up.  Look around for some UV mapping tutorials, check rlk's comments above on Neverball materials vs Wings3D materials and maybe start by tweaking a simple misc_model in a map instead of ball.

71

Re: A new kind of ball

I'm keen on making some balls too. I'd love to try my hand at a sun with glow and animated solar flares (should be doable with resizing sprites the same way waves are done).

Maybe an entire solar system set could be cool. I'm definitely loving Saturn (though the pendulum effect isn't good for it, I don't think).

Cheese
==========
cheesetalks.net

72

Re: A new kind of ball

Yes paxed, it sounds like your OBJs have no "vt" lines.  These represent UVs.  The easiest way to produce UVs in Wings is to use Image Snap mode.  You start by using "File/Import Image..." to load a texture image.  Then select "Tools/Snap Image".

Now you're in snap mode and the texture is overlayed on your view of your object.  Select the faces you want textured, and position them with the camera controls until the texture fits as you like.  Then, in face selection mode, the first option of the popup menu is "Snap Image".  This applies the texture to the selected faces.

Now, when you export that as OBJ, mapc will recognize those faces as having UVs and will include them.  For some materials, like glass and chrome, the UV mapping doesn't matter, but the UVs do have to be there anyway.

73

Re: A new kind of ball

Thanks, RLK, that works!

big_smile

74

Re: A new kind of ball

Can't wait to see some of the things paxed thinks up! Maybe a "lava ball"...

75

Re: A new kind of ball

Here's my first "real" try.  Not quite happy with it, but damn this stuff is exciting.

http://parasti.googlepages.com/spikes.png