Archived

This topic is now archived and is closed to further replies.

I need some help starting this...

This topic is 5000 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Hey everyone... I''ve been very interested for years now in game programming. I have done a few little things like make teris clones, and breakout...and I also made a 2D scrolling tile map with an animated character walking around on it.... But now I feel that it''s time for me to move on and get serious about game programming. I want to start learning 3D basics, and then move to advanced stuff. I''ve been going through the nehe site, and reading little pages here and there, and now I''m starting to incorperate these things into my engine, however I just need a little help with starting. I read people here talking about skeletal animation, and ROAM rendering techniques, and I wish I could join in but to be honest, I''m just at the very beging of 3d programming and am not ready for those topics yet. Can you guys post little stories, or ideas about your first experiences in starting out? I''d like to know what others did to break into 3D development. Thanks, and sorry this post is so long... I would just like some advice Michael Rhodes

Share this post


Link to post
Share on other sites
I had a lecture at my university about the maths stuff in opengl, so I already knew about these 4x4 matrices. Then I just found the NeHe pages, did some tutorials and tried my own stuff.

After some while, I liked to do some more things which are not explained in the tutorials like dot3 bumpmapping and global terrain rendering. Google was my friend here. There is always someone who has already implemented what you need and has - by chance - put up a very good tutorial and/or some paper explaining his method. Just enter something + "tutorial".

So my advice: Just go on "project-driven". Just think about what you want to do, then try to figure out how to do it and google for it. Read tutorials, papers and extension specifications. Gradually, you will learn all that stuff like using extensions and vertex/fragment shaders. At least that works for me. I think, there''s no way in first learning everything and then applying everything. Learn a bit, apply it, learn something more and so on. By the way, that''s much more fun.

I''ve set up a terrain rendering engine by just plugging each new bit I learned into the engine. The result is a HORRIBLE design, but I''ve learned a lot!

Share this post


Link to post
Share on other sites
Ahhmm..OK I''ll try to remember...

A few years ago...maybe 7 or 8 I was in the same situation.
In that time there were not as much tutorial-sites and that stuff (NeHe & co.). So I started with simple things like 3-DOF voxel-rendering, which can be inserted into almost every 2D-engine or rendering-library. This gave me a first grasp of dealing with coordinate systems and all that linear stuff.

The next step was writing alot stuff regarding software rasterizers. (Remember, there were no usable/afordable hardware-renderers on consumer-platforms at that time)

It took me quite a time, from starting with naive implemented triangle rasterizers to full-fledged bumpmapping-phong-shading-rasterizers that were reasonably fast. But every small game I made and every technology-testbed I wrote took me a step further to my goal - being a professional.

And today, I am coding surgery-simulators with a lots of open-research, like wire-simulations that let you make knots with two tracked virtual surgery-tools, and all that stuff.

So my hint is....straight keep on coding....even if your improvement is as small as can be...sooner or later, you will reach your goals, as I did.

Have a load of fun.


Share this post


Link to post
Share on other sites
Thanks for the replys guys In the time it took you two to reply I started by making a very simple Mesh structure, and loaded it with 2 polygons (triangles) and worte a simple loop to render it in OpenGL I will still have a lot to do to make this better, but it''s a start. I really don''t want my engine using imediate mode for to long!

Here''s a question... do you store everything in a mesh structure / class? I mean things like normals, tex coordinates, colors, etc? Or would you have a couple different structures to store such things?

Thanks,

Mike

Share this post


Link to post
Share on other sites
To start:

Write a simple buffer-class that is capable of resizing itself and that makes it easy to append data and so on.

After that make your rendering code as flexible as possible...the concept used by D3D, I mean vertex declarations are a good sample for flexibility.

Then you can begin building meshes, by simply generating your data and building index-lists on the fly. It is pretty straightforward to use and comes in handy, if you have to, for example,add tangents, normals, more tex-coords or other stuff into existing meshes.

In a nutshell, this means, you should be capable of telling your renderer, where to find data, and how to interprete it...the you have a very comfortable framework at hand, that makes furter development as easy as stamping on kittens*g*!

Regarding above statement: Does anyone remember Duke Nukem?

Share this post


Link to post
Share on other sites