Will anyone else be doing the Global Game Jam?
[url]http://globalgamejam.org/[/url]
[url]http://globalgamejam.org/Locations[/url]
[QUOTE=slime73;34152721]Will anyone else be doing the Global Game Jam?
[url]http://globalgamejam.org/[/url]
[url]http://globalgamejam.org/Locations[/url][/QUOTE]
Probably. I went to Nordic Game Jam last year, and I will most likely go this year too.
[QUOTE=slime73;34152721]Will anyone else be doing the Global Game Jam?
[url]http://globalgamejam.org/[/url]
[url]http://globalgamejam.org/Locations[/url][/QUOTE]
Maybe, I did the last two. Depends how much work I have to do around then
[QUOTE=FoohyAB;34030304]Haven't posted in a while, mostly lurking.
Anyways for the past bit or so I've been working on making my own (very simple, I'm not planning to ever use it) engine, using only the default SFML 2 functions and make everything else myself.
Recently I got fullscreen shader effects done! (They're really bad, I'm not exactly a super wiz at GLSL fragment shaders)
[vid]http://dl.dropbox.com/u/1179448/Misc%20Video/pong_shaders.webm[/vid][/QUOTE]
Where did you get the music while playing?...
[url=http://code.google.com/p/mochalua/]mochalua[/url] is pretty much the best Java implementation of Lua ever. They neatly carry over names from the Lua C API over to a single class called LuaAPI. I don't have to learn any new API or some weird shit convention of creating a Lua state or making a library or new pseudo-type for Lua, or manually create a closure (what the fuck?).
I'm toying around with Minecraft to just quickly see if I can create a Lua state in it. Not even load a script or run a string - I just want to know how to set up my development environment to use it.
[editline]10th January 2012[/editline]
[QUOTE=Rayboy1995;34152990]Where did you get the music while playing?...[/QUOTE]
[media]http://www.youtube.com/watch?v=dclghhnW4oU[/media]
-snip- This isn't the right place, sorry.
In the process of writing a CHIP-8 emulator using the fantastic documentation here: [url]http://devernay.free.fr/hacks/chip8/C8TECH10.HTM[/url]
Images of my progress so far:
[i]A brand new project[/i]
[t]http://i.imgur.com/WnKom.png[/t]
[i]Loading a test file into RAM[/i]
[t]http://i.imgur.com/A9jem.png[/t]
(At this point I found some public domain ROMs for CHIP-8, including the ever-popular maze generator, which I chose to be the first ROM for me to emulate.)
[i]Getting the first sprite up on the screen[/i]
[t]http://i.imgur.com/VACra.png[/t]
[i]Fully emulated![/i]
[t]http://i.imgur.com/L4HQ0.png[/t]
Today I realized that my laptop had no sleep button.
Lots of Google and a little batch later, Ctrl+F12 is my new sleep button.
It might be the most useful thing that I've ever made.
I got a planet orbiting a star.
[img]http://img515.imageshack.us/img515/1435/sunplusplanet.png[/img]
I might have some issues with figuring out what's the best scale for everything.
[QUOTE=amcfaggot;34153088][media]http://www.youtube.com/watch?v=dclghhnW4oU[/media][/QUOTE]
Holy crap. I've never played VVVVVV but that song is so awesome I really want to now.
Was the same way with SMB, actually. I loved the music in that game, ended up picking it up, and it became one of my favorite games.
I'm a sucker for a game with good music...
[QUOTE=dvondrake;34153858]Lighting is starting to come together.
[img]http://i.imgur.com/GZf1F.png[/img]
Different materials can have different specularity, and lights can have different colors and radii.
Currently have specular factors and specular lighting on their own separate render targets, but that's only temporary. Going to put specular factor in the alpha channel of the normals, and specular lighting in the alpha of lighting. Currently debating whether I really need to leave out the normals' Z, as I don't think I'll really be using that extra channel it frees up. Is the memory still allocated regardless if I'm using it, or does leaving off Z still save me some memory?
Next up: Loading albedo, normals, and specularity from texture maps. Then shadows. Then all the "standard/must-have" stuff for a rendering engine is pretty much there and I can work on more advanced and cool stuff.
[b]Edit:[/b] Fixed my lighting formula, takes into account proper light decay now. Image has been updated, [url=http://i.imgur.com/jS71m.png]here's the old one[/url] for comparison.[/QUOTE]
Z will still be fully accounted for as 1, so whether you render to the blue channel or not doesn't matter. But I would keep normal Z because it's actually useful, for both post processing and later on (especially if you convert over to use tangent, bitangent and normals for tangent-space normalmapping, you'll need to have 3 3 channel floats to compose into a matrix to transform your tangent normals by for correct lighting.)
Tried my hand at pool table detection for eventual recognition of the individual balls. Gonna need the whole table in the shot I think.
[img]http://i.imgur.com/YGP6d.png[/img]
[QUOTE]recognition of the individual balls.[/QUOTE]
:quagmire:
[QUOTE=Xeon06;34155586]Tried my hand at pool table detection for eventual recognition of the individual balls. Gonna need the whole table in the shot I think.
[img]http://i.imgur.com/YGP6d.png[/img][/QUOTE]
Somewhat related: I wondered if you could make 'virtual' pool that had a camera + projector projecting down onto a pool table, then track the pool cue with the tip and a difference from the blank pool table. Then spawn balls, calculate the force of movement and if a pool cue collides, etc.
Alternatively you could use it to assist you - track the balls and figure out the best shots and what angle/power you should hit the ball with and then project that. Would be a fun project.
[QUOTE=Lord Ned;34155896]Somewhat related: I wondered if you could make 'virtual' pool that had a camera + projector projecting down onto a pool table, then track the pool cue with the tip and a difference from the blank pool table. Then spawn balls, calculate the force of movement and if a pool cue collides, etc.
Alternatively you could use it to assist you - track the balls and figure out the best shots and what angle/power you should hit the ball with and then project that. Would be a fun project.[/QUOTE]
It's been done. If I remember correctly, someone even posted a video of it here. The projector showed the line of trajectory, angle of impact and such.
I'm extremely pleased with the amount of progress I'm making. I've already finished remaking my sprite loader, and the map loader is underway
[img]http://dl.dropbox.com/u/29407915/luajit2.PNG[/img]
Converts now properly tables, entities, vectors, angles and pretty much the rest. GLuaCall will call the GLua code and gives back the result to LuaJIT, only when a result is expected of course.
However this adds the ability to call things like SetPos on objects and such.
Trying to use an image file to draw each floor (it's a dungeon crawler) of the game. Barely fucking works so far.
[QUOTE=Zeh Matt;34156334][img]http://dl.dropbox.com/u/29407915/luajit2.PNG[/img]
Converts now properly tables, entities, vectors, angles and pretty much the rest. GLuaCall will call the GLua code and gives back the result to LuaJIT, only when a result is expected of course.
However this adds the ability to call things like SetPos on objects and such.[/QUOTE]
I'm curious as to how you're dealing with GMod's native type extensions, as well as typical userdata.
[QUOTE=amcfaggot;34156612]I'm curious as to how you're dealing with GMod's native type extensions, as well as typical userdata.[/QUOTE]
References
[QUOTE=Zeh Matt;34156619]References[/QUOTE]
That doesn't explain two Lua states running concurrently; one which doesn't know anything about the other's bindings, as well as being a different dialect.
I was hoping to do a generic unbox method but apparently it cant tell what return type you want.
[csharp]
string str = Unbox(obj); //It cant tell that I want a string? :(
static T Unbox<T>(object o)
{
return (T)o;
}[/csharp]
Was hoping to save a bit of typing. Anyways this is just me thinking of ways around the lack of fieldof() in C#. So instead of defining [Order] on every field and having it get lost in the hierarchy I thought about defining a list of setters/getters using lambdas.
[csharp]
public class Fields<T>
{
public Action<T, object> Setter;
public Func<T, object> Getter;
public Fields(Action<T, object> setter, Func<T, object> getter)
{
Setter = setter;
Getter = getter;
}
}
public class Bot
{
public string Name { get; set; }
public string Id { get; set; }
static List<Fields<Bot>> Fields = new List<Fields<Bot>>
{
new Fields<Bot>((b, o) => b.Name = (string)o, b => b.Name),
new Fields<Bot>((b, o) => b.Id = (string)o, b => b.Name)
};
}
public class Fields2
{
public Action<object> Setter;
public Func<object> Getter;
public Fields2(Action<object> setter, Func<object> getter)
{
Setter = setter;
Getter = getter;
}
}
public class Bot2
{
public string Name { get; set; }
public string Id { get; set; }
List<Fields2> Fields;
public Bot2()
{
Fields = new List<Fields2>
{
new Fields2(o => Name = (string)o, ()=> Name),
new Fields2(o => Id = (string)o, () => Name)
};
}
}[/csharp]
Second method is a bit better because not only is it less typing but you can incorporate it into a base class more easily. Only downside is it adds a list of fields to every object :(.
So I've been working on some android game for awhile now (not that I've put tons of time into it, it's just something I've been working on off and on. Mostly off.)
Anyways I never had a theme for it and it was just "colored squares" for sprites and stuff. Finally my friend who I've been working on it with me came up with the theme of snowballs with faces. Anyways I decided that static snowballs with faces was boring (obviously) so I made a key frame animator thing for their eyes so that they aren't as boring. The animator is really limited (only does their irises) but I threw it together quite quickly. It then renders the animation as a png file when you're done.
Have some videos:
(The key frame animator, all java applet stuff)
[media]http://www.youtube.com/watch?v=HyR2Xuj9MN0&feature=youtu.be[/media]
(An animation running on an android tablet, focus sucks but I'm too lazy to record again)
[media]http://www.youtube.com/watch?v=Gz5TKuvka8k&feature=youtu.be[/media]
[QUOTE=Smashmaster;34154967]I got a planet orbiting a star.
[img]http://img515.imageshack.us/img515/1435/sunplusplanet.png[/img]
I might have some issues with figuring out what's the best scale for everything.[/QUOTE]
Reminds me of the wormholes in EVE.
[img]http://wiki.eveonline.com/wikiEN/images/e/e3/Wormhole_-_1.PNG[/img]
Working on arrays at the minute, ironed out a few issues as well - you can do stuff like this now:
[img]http://i.imgur.com/PaJ4A.png[/img]
Passing a procedure as an argument to a procedure! :D
Working on getting parallax back/foreground scrolling working.
[QUOTE=Xeon06;34155586]Tried my hand at pool table detection for eventual recognition of the individual balls. Gonna need the whole table in the shot I think.
[img]http://i.imgur.com/YGP6d.png[/img][/QUOTE]
That's a pretty cool idea - you making a `best shot` app?
[QUOTE=dvondrake;34153858]Currently have specular factors and specular lighting on their own separate render targets, but that's only temporary. Going to put specular factor in the alpha channel of the normals, and specular lighting in the alpha of lighting. Currently debating whether I really need to leave out the normals' Z, as I don't think I'll really be using that extra channel it frees up. Is the memory still allocated regardless if I'm using it, or does leaving off Z still save me some memory?[/QUOTE]
It's generally not the space that is the issue, it's the fact that normals are one of the most important values in many of the lighting equations, and any extra precision you can get helps a ton. Storing only X/Y enables you to use 16-bit channels and still jam both component into a single 32-bit attachment.
Consider specular highlights. You're raising the dot product of the normal with a lighting/view vector to some power. This means that, if you have a large-ish exponent, small variations in normal value can create large changes in lighting intensity. With 8-bit RGB normals you can easily run into quantization artifacts.
[QUOTE=Legend286;34155245]Z will still be fully accounted for as 1, so whether you render to the blue channel or not doesn't matter. But I would keep normal Z because it's actually useful, for both post processing and later on (especially if you convert over to use tangent, bitangent and normals for tangent-space normalmapping, you'll need to have 3 3 channel floats to compose into a matrix to transform your tangent normals by for correct lighting.)[/QUOTE]
Tangent-space to view-space transformations happen [i]before[/i] normals are written to the g-buffer...
And the whole idea is that you can get the z-component back anyway. Because everything lies on a unit sphere and any point on a sphere can be fully described in two values, z-information is mostly redundant.
Sorry, you need to Log In to post a reply to this thread.
