I can easily write a program which loads 3d models and UV textures, but skeletons and vertex keys! I don’t know about that.

hrmm, not talking about using actual 3d models. I was thinking cel shaded 3D rendered out to BMP or another file format. As for cel shading, it’s all about the artists and the textures applied. you can get detail as long as the texture still carries out the handdrawn look. but anywho, i’m available if you guys need an artist and choose to use the 3d rendered to image file technique i mentioned above. actually, why doesn’t someone throw me a character idea and let me see what i can come up with to prove that this is a viable technique?? any takers?

A kid in shorts with a bassball bat.

seantree, if you produce a spritesheet, I can release a demo with the character for everyone on this thread to try.

Hah, I got the same impression.

To me, it seems much easier to write a program that will take a 3d model and export pre-rendered sprite sheets than to expect people to draw a sprites 1000+ frames consistently with quality. So in the game, you’re still dealing with sprites, but theres an external program we’d have to make that will take the model, run it through all the animations it has, and create sprite sheets out of that.

That seems like the fastest/easiest way to do this. The talk of cel-shading is just to give this an old-school 2d flat look that we all like, I would think…

Anyway, I’ve been looking over the source code, pretty good stuff. If I were to load this into Visual 6.0 and compile, anything special I need to set up with (I assume OpenGL and stuff)?

Street Fighter 4: SRK MEETS THE STREET

I’d play it.

This sounds like the best way to go about it.
Also I agree, Cel-shading is only for the traditional handrawn look and if it is going to be just more work I would just dump that idea.

While the code has been written from the start to run on both Linux and Windows, I have up to this point only been compiling it with gcc. You will need to know how to link and compile SDL apps with OpenGL in order to compile my application. So you need the SDL libraries, including SDL_image and SDL_mixer.

I really need to check out the SDL libraries. I have been using Allegro because it is so easy, but I think I want something a little more powerful.

Haven’t the SDL libraries been ported over (at least in a limited way) for the Dreamcast as well?

So basically we are “capturing” 3d models just like the captured video of real people doing moves in the 2d mortal combat series.

I think this is worth experimenting with. If the outcome is shitty looking then we can abandon it, but the smaller workload is worth the time spend to test it, I think.

I have seen a mugen version of Jin Kazama that worked off screenshots from Tekken 3. I think that is a really shitty version of what I am thinking of.

EDIT: Someone get Zinac some sprites!

This may sound like a wierd idea, but let me know what you think. What would you guys think of an amature game without actually human like detailed sprites, but with sprites figured with simple shapes and colours? I can’t quite explain what I mean, like you don’t even need detaled sprites, as long as the hit boxes are there and are solid that’s all that matters. The sprites could be shilouette like made out of different shapes, colours and hit boxes to differentiate them from eachother. I’m gona shut up now or I’ll probably just repeat my self.

I could make the sprites anyway. If anybody want to.

In my opinion this would give the appearance of the 2D MKs, or KI.
Also, has anyone mentioned they can actually animate 3d characters?
I think it’s much more difficult to animate a 3d model as an amature if you’ve had no experience than someone who can draw well and consistantly draw animations. Just because you can put together a character mesh, does not mean you can rig and animate it properly.

Also keep in mind that professional game developers have motion capture devices, and a majority of animation in games now are developed using motion capture.

i’ve got animation skills using 3ds max biped or built in bones. Give me some time and i’ll throw something out. My new job really eats my time up so as soon as i have something to show i’ll post it up. As for mocap, sure it can save some time, but even the big dev studios go back and refine the mocap data by hand. The amount of anims a fighting game character would need should not be that high. Idle, walk forward/back, crouch/ jump, standing/crouching attacks, specials, and throws. once one char is finished i could take the data for that character and apply it to the next chars rig, with a little tweaking we’d have an all new set of anims.

I’m a scripter. Can adapt to mostly any language, if you guys need help hit me up.

I want to note that anyone thats has an IRC client can visit the development channel #fgdev on Efnet

I’m currently rewritting the input handling for my engine. Before I lay down any new code I would like your input first.

I was planning on people showing up to the IRC channel to discuss development, but since thats not working out I’ll just post technical information here.

I am going to start with a few assumptions to design the algorithm to handle normals, specials and supers.
The notation I will use is as follows:

  • is used to indicate whether one or two buttons are held down for a tick. Example: A+B means either A or B or both can be held down for the statement to be true.
  • is used to indicate logical AND. A * B, also can be written AB.
    for the statement to be true, button A and button B must be held
    for the tick.
    ! attached to a button means the button must not be pressed for the tick. Example: !AB means B must be pressed and A can’t be.

I will use the SRK icons for the directions, and letters to indicate buttons.
:u::d::l::r:

!:u:B

I will also use commas ‘,’ to indicate execution of specials and supers
Example: :d:,:d::r:,!:d::r:,A

Diagonal arrows are not used since most controllers only have 4 switches, and I want to indicate the digital state of the switches as precise as possible.

I’ll return a little later with the specific way in which I want to handle normals, specials and supers. Feel free to correct/improve my notation.

Edit: Continuing

Now I will explain how commands are detected.

Directions
As simple as when a direction by itself is pressed the program executes the designated action if nothing is preventing it to.

Normals

Normals consist of either just an attack button, say A, B, or C, or an attack button and a direction (:u:).
Examples: :r:A ; C; :l:AC

The program executes the normal when it first detects a normal input, if there is not some reason thats preventing its execution.

If anyone is interested, I’m a professional sprite artist. If you guys want to go in the 2D sprite based way - I could help, HOWEVER, I am extremely busy at the moment working on two projects, so the only way I could help is with minimal sprite size and animation (ZOMG DAS 2 FRAME PUNCH). However, with any engine, the amount of animation doesn’t matter so much (CvS2). If you guys are interested, the offer is here.

If anyone is a good artist, even drawing a few frames (a two-three frame punch) and scanning it would allow me to convert it into a workable sprite (workable size + 16 color max limit … unless you guys wanted to go 32 colors and beyond ala SFIII).

Keep in mind I do sprite work for handhelds (GBA/DS and all mobile … well all the work is there anyway);

Even if we get one character done this way, that’s already an awesome start, and we could start experimenting with the engine and stuff.

By workable I mean something of this size;

http://img.photobucket.com/albums/v176/Sonic_Reaper/ula11.gif

And a select screen portrait;

http://img.photobucket.com/albums/v176/Sonic_Reaper/ula12.gif

Wow that is very nice Sonic_Reaper, I’m impressed.

nice work reaper. looks great