superh posted:

That looks like it's in the sprite editor, is that what you mean? You can definitely script it. I'm not at my desk right now but, I know I grabbed a plugin for parsing texturepacker sprites and that's doing it, if you need an example.

Joda posted:

I managed to implement an alias map that welds together two vertices if they are exactly the same. It's entirely appropriate for the coding horrors thread, and I expect to hate myself whenever I have to revisit it, but at least it works. Still trying to figure out how to adjust for angle.



It can't handle margins of error (since I use a simple hash map,) so the vertices have to be exactly the same, so there are still some artifacts here and there.

I then proceeded to follow wayfinder's advice to fix my quad problem; still not defined but getting there. Video will have to wait.

Thanks a bunch for the help!

AntiPseudonym posted:

Yeah, that's pretty much exactly what I'm after. If you could post an example that'd be brilliant!

superh posted:

It looks like the relevant code is spread all over the place in my project, but the guts of it are:

Create a List of these SpriteMetaData's, these are basically the rects you want:
http://docs.unity3d.com/ScriptReference/SpriteMetaData.html

Then you take that List and apply it back to the sprite via the texture importer.

code:

TextureImporter texImp = AssetImporter.GetAtPath(path) as TextureImporter;
texImp.spritesheet = sprites.ToArray();
texImp.textureType = TextureImporterType.Sprite;
texImp.spriteImportMode = SpriteImportMode.Multiple;
AssetDatabase.ImportAsset( path, ImportAssetOptions.ForceUpdate );

Actually building the SpriteMetaDatas happens elsewhere but it's fairly straightforward.

Hope that helps!

LookingGodIntheEye posted:

What's it like to work as a playtester?

AntiPseudonym posted:

Although if you have aspirations of using it as a stepping stone into the industry, and don't mind the actual work of methodically playing the same tiny fraction of a level over and over again to reproduce a rare bug that happens 1/100 times, it can be a fun job if you work for the right studio.

LookingGodIntheEye posted:

This is what I'm looking for. I'm just beginning my programming career at this point and I'm open to exploring multiple fields.

LookingGodIntheEye posted:

This is what I'm looking for. I'm just beginning my programming career at this point and I'm open to exploring multiple fields.

Police Automaton posted:

I don't really know where else to ask the question I have so I am just going to ask here.

I'm from an electronics/programming background (lately more electronics than programming because I switched careers from being self-employed) but like every aging nerd I always had that dream of writing my own vidya. Over the years I always had a brainchild which once started off in the ancient days of the old Commodore Amiga and sort of progressed with platforms, programming languages and my own knowledge. The latest incarnation I shaped in C# (not utilizing any 3rd party engine because they didn't exist then, but just pretty much from scratch utilizing OpenGL when everyone still said C# will never be a thing for video games) just writing a very basic 2D engine and just exploring a few ideas and never tying it into a coherent package. (for example writing a multithreaded pathfinding engine, or a fractal landscape generator)

I would like to finally make the push and realize my idea into a coherent game for others to enjoy. I just really notice again and again I do not really know where to start organizationally and feel a bit overwhelmed regarding the options. Designing a game is very different from everything else I have ever done before, and nowadays with all that steam stuff 3rd part engines everyone seems to employ and etc. I am a bit overwhelmed what the best path to take would be to take idea and code to something that people actually can end up playing. I don't want to turn into the dwarf-fortress guy. If anyone has a primer that outlines organizational strategies of game development I would be much obliged. I don't plan on getting rich (or even quitting my current job for that matter) but I would like to turn my idea into something more "real" in the next few years and into something other people could also play. I hope this wasn't too outlandish a question and makes some sense.

LookingGodIntheEye posted:

This is what I'm looking for. I'm just beginning my programming career at this point and I'm open to exploring multiple fields.

Centripetal Horse posted:

Program a game. Seriously, make anything, as long as you complete it. There really is no substitute for experience. Writing a game from scratch will lead you into a lot of roadblocks, and expose a lot of questions and challenges that I think will help give you a starting reference for your larger project(s). You can make an effort to include things you expect to be in your big game, but even if you don't, you will learn a lot.

If you're not familiar with Handmade Hero, definitely have a look: HandmadeHero.org. Casey Muratori is programming a game, from scratch, live on the internet, using no third-party libraries. The forums there are becoming pretty active, and some fairly well-known programmers, from game-industry to language-designers, participate.

Start haunting the Indie Gamer forums and other message boards dedicated to the industry. Like most things, there is a lot of lovely advice and information out there. Some people will say that's what I'm giving you, right now. Do your best to farm the bits that apply to you and your goals.

I am sure other people will be along with more specific advice, shortly. Try not to get too hung up on how other people do things, though. In my experience, the process is so unique to the individual that, no matter what people tell you, you're going to have to discover about 75% of it for yourself.

Onion Knight posted:

Police Automaton posted:

Basically, this was all I really wanted to hear I think.

Police Automaton posted:

I for myself see no use in using any 3rd party engine as I dusted off my old source with all the interface classes and it still compiles fine (without changes!) and works both in Linux and Windows, and it does all I need for now. I remember putting a lot of effort into only updating screen regions that had changes (my computer really was that slow back then) and this nets me five-digit framerates now on my current computer which cracked me up. I do think I should learn how to get better organized as I tend to get lost in details and then never manage to get anywhere. But that's not really a game-development specific thing to learn. It's a shame that I can't even do the most basic pixel art to save my life.

Yodzilla posted:

I could be really, incredibly wrong but isn't four rays a frame pretty trivial performance wise? I'm working on a 2D action game that uses the prime31 CharacterController2D and it uses a bunch of rays to detect collision every frame and it runs just fine with multiple players at once. In bigger games aren't raycasts used all over the goddamn place for checking any number of things?

Onion Knight posted:

I have a perspective projection camera I want to track a point with as it moves around a rectangular stage. The stage is 2D (along the X/Y axis) and the point (the centroid of player positions) has a constant Z.
However, I want to prevent the camera’s view from extending past the level bounds. This would be trivial with an orthographic projection, but I’m having trouble with a perspective camera (especially because it can pivot along its Y axis, in certain situations)

The obvious brute-force solution would be to raycast and stop the camera from moving when the rays hit one of the boundaries (or from one of the boundaries?). This would be at least four rays a frame, every frame, and seems way too expensive.
I don’t necessarily want the camera to move too much, or too quickly, so I was going to have it gently lerp towards the point. I might be able to get away with raycasting to/from one of the four bounds every few frames, then adjusting the centroid (and thus the camera path) based on that. This would probably work, but it might make the camera feel too loose, and might be too expensive anyway
The other approach I thought of might be better, but I’m not sure how to implement it yet, but it would be defining a bounding box for the stage, then calculating the camera frustum slice at player-z, and making sure none of those points intersect. I should only have to recalculate if camera changes rotation (and even then I’d only have to determine the frustum at the end rotation).

Is there an easier way to do this? Or a best practice for keeping the viewport in within a bounds?

Unity, if it helps.

Bondematt posted:

Yeah, 4 raycasts an Update is nothing. On PC I'm running a couple hundred just to make my mesh for a stencil buffer shader every Update.


If you know the Bounds X and Y you want the camera limited to; couldn't you just take the Camera's current position and then stop it when it hits whatever X/Y you want it to stop at?

Yodzilla posted:

I could be really, incredibly wrong but isn't four rays a frame pretty trivial performance wise? I'm working on a 2D action game that uses the prime31 CharacterController2D and it uses a bunch of rays to detect collision every frame and it runs just fine with multiple players at once. In bigger games aren't raycasts used all over the goddamn place for checking any number of things?

Bondematt posted:

Yeah, 4 raycasts an Update is nothing. On PC I'm running a couple hundred just to make my mesh for a stencil buffer shader every Update.

quote:

If you know the Bounds X and Y you want the camera limited to; couldn't you just take the Camera's current position and then stop it when it hits whatever X/Y you want it to stop at?

Onion Knight posted:

Oh, huh. I've only ever heard "raycasts are expensive" with no qualification so I've been kind of gun shy about using them. I think in that case I'll just try the brute force approach, and if that doesn't work (or I get a big performance hit) maybe then try to figure out frustum math. Thanks for the help.


If the camera was orthographic, sure. I mean, I could still do this, but I want to maximize the gameworld on-screen. If I sit the camera, physically, on the bottom of the level I still have half the screen filled with floor.

Onion Knight posted:

I have a perspective projection camera I want to track a point with as it moves around a rectangular stage. The stage is 2D (along the X/Y axis) and the point (the centroid of player positions) has a constant Z.
However, I want to prevent the camera’s view from extending past the level bounds. This would be trivial with an orthographic projection, but I’m having trouble with a perspective camera (especially because it can pivot along its Y axis, in certain situations)

The obvious brute-force solution would be to raycast and stop the camera from moving when the rays hit one of the boundaries (or from one of the boundaries?). This would be at least four rays a frame, every frame, and seems way too expensive.
I don’t necessarily want the camera to move too much, or too quickly, so I was going to have it gently lerp towards the point. I might be able to get away with raycasting to/from one of the four bounds every few frames, then adjusting the centroid (and thus the camera path) based on that. This would probably work, but it might make the camera feel too loose, and might be too expensive anyway
The other approach I thought of might be better, but I’m not sure how to implement it yet, but it would be defining a bounding box for the stage, then calculating the camera frustum slice at player-z, and making sure none of those points intersect. I should only have to recalculate if camera changes rotation (and even then I’d only have to determine the frustum at the end rotation).

Is there an easier way to do this? Or a best practice for keeping the viewport in within a bounds?

Unity, if it helps.

The Atomic Man-Boy posted:

I'm just getting into building my own game.

I count myself a decent programmer, and know a fair bit of unity. Its going to be an interesting voyage. Sadly, I know nothing making art assets.

I'm trying to make a hack'n slash type of game. But I don't know about the modeling/animation/rigging/skinning pipeline. I want to whip up a skeleton body in Blender that I can use as a placeholder for development, which a real artist can expand upon later. Here's the requirements:

1) A human shaped skeleton that can be used for both male and female characters meshes.
2) Can be used for all animations, i.e. "sword slash 1", "walk", "run", "die" etc.
3) Can be re-skinned programmatically with a few different faces.
4) Has a number of sockets that I can add armor/clothes when I have an actual artist.

Unfortunately, as I said, I don't know anything about the asset creation pipeline. So I'm not sure how to go about this. I need to know how to build and animate a skeleton, and use a placeholder that wont be an impediment to my future artists and wont conflict with my unity project when I replace it with skinned/clothed models. Any advice on how to do that is greatly appreciated.

floofyscorp posted:

There are lots of character models and animation packs on the Asset Store that you can use for placeholders, or you could try something like MakeHuman to make a more customised character model. With the ubiquity of Mecanim it's not terribly hard to find bipedal animations that you can use on any relatively standard human-shaped rig, but if you're thinking about a less than standard-shaped character you'll likely have to make everything for it from scratch. Personally I'd feel that creating and animating a skeleton first, and retrofitting a character model later, is a bit of a backwards workflow - normally I'd make the mesh first, then create a skeleton that fits it and animations that work with that etc. If you're going for standard human proportions though, I guess it's not necessarily a big deal. Things like having an allowance for different faces and armour etc are not something I'd worry about at the skeletal level; that'd be all in the mesh.

Japanese Phone Box posted:

So I'm dicking around with freeglut as part of my university's computer graphics classes. I'm trying to program a 2D vector based game in the style of Vectrex games in C++ and I'm doing fine for now, but I have two important questions (keep in mind I consider myself a beginner in C++):

1- How do I structure my program in a decent way?

ErIog posted:

The solution is not to use SpriteRenderer for something like that. It seems intuitive that you should, but getting it to work the way you want it with a SpriteRenderer is going to involve you writing graphics code to make it work.

The standard answer to this question in Unity is to use textured Quads with AlphaBlended shaders that use Textures set to Wrap Mode: Repeat. You might have to futz with some values to get Unity to map the pixels to world scale properly, but once you do that it works fine.

http://answers.unity3d.com/questions/587007/im-trying-to-create-a-tiled-sprite-what-am-i-doing.html

ShichiNoBushi posted:

Thanks, your answer helped... sort of. Though now if I want to make multiple distinct objects to create the environment, I need to make a new Material with it's own X-Y Tiling ratio based on the size of the object I'm creating.

Bob Morales posted:

You can read Game Programming Patterns for free online - http://gameprogrammingpatterns.com/contents.html

ErIog posted:

You should only need to make a material for each sprite texture. I didn't have problems with multiple sizes when I used that solution. Things just tiled the texture as much as they could along the length/height of the object.

ShichiNoBushi posted:

What I noticed trying to get it to work is that if I wanted to make an object 5x1 tiles large, I'd first need to make the material tile 5 X and 1 Y. Then if I adjust the object to the size I wanted, it would be tiled appropriately. Though I only did this with one object, I would expect that If I wanted a different wall 1x5 in size but used the same 5x1 material, the texture would be stretched like before rather than in a 1x5 grid of appropriately scaled sprites. If possible, I would like to use a single material for each texture I use, so if you can more specifically tell me how to accomplish that, I would appreciate it to save on needlessly complicated and resource intensive methods later.