Josh

What language are you writing this in?

Yes. The first image looks terrible. Maybe his DDS saver is using a different technique than whatever you used?

The evaluation kit does not include the FBX to GMF converter, so the model viewer can't load your file.

This might be feasible if it is limited to movement/creation/deletion and terrain, and not stuff like undoing the last script change or crazy stuff like that.

As mentioned, official documentation will come to be on the site, in addition to the wiki.

Well who pressed the button?

Fascinating. You may be right.

I would not use Lua for the base-level AI, for a complicated game. I would use it to handle AI reactions and changes in behavior, but not for any math intensive stuff. For example, you might make the bot change from "sad" to "angry", but you would use machine code to handle those actual behaviors. For other less complex games, Lua will do just fine for simple AI.

The second thing you said.

Directional light shadows are designed to light an area of kilometers. You are not going to get high precision shadows on something as small as character facial features, with this scale of lighting. You can decrease the directional light shadow distance to have more detail in a smaller area, but you will lose shadows in the distance. You can also increase the light's shadow map resolution. This problem is common to any real-time engine that uses shadow maps. Here is an example from Crysis:

A mesh oriented to face the camera.

Add UpdateAppTime() to the loop.

Edit the particle vertex shader so rotation is not used. I am not sure why you are using an emitter for this, though. In version 2.31, the particle velocities will be sent to the vertex shader. This could be used to orient the particles in the direction they are moving, for sparks and things like that.

-Add your normal map and specular map to the material. There are lots of examples in the SDK of materials with a bumpmap. -Disable bloom and HDR. -Use a darker ambient light color.

For that particular scene, a nav mesh might be overkill. I do want to integrate the recast library into the engine, but it's definitely not necessary for something like this.

Version 5 files use twice as many bytes per pixel, so they have higher resolution / less lossiness.

The framework class should detect the resolution change automatically and recreate the buffers on its own.

That game is fun but extremely annoying. They don't give you enough money or time to build anything effective.

Do NOT use DXT1, ever!

That's exactly what we're doing. We have the ability to add various feeds into the html pages, so the whole site will be more interactive and interesting.

Increase the camera near range.

No, you would not have enough video memory or texture bandwidth. Directional lights are very specialized because you cannot possibly make a texture big enough to shade the whole scene, with a decent resolution. The whole challenge of directional lights is to use more pixels in the foreground and fewer in the background. The two main ways of doing this are to have several renders of the scene at different resolutions, or to render the shadow with a skewed matrix that uses more pixels closer to the camera. There's little I can do to optimize directional lights, as they are very much a brute force technique. Far Cry 2 seems to use a skewing technique that doesn't seem to have visual errors, but of course they are going to go to great lengths to hide any errors their technique produces by not using certain sun angles. If I can devise a decent skewing technique, it could mean one shadow render of the scene instead of three, but I am not sure I have ever seen an error-free implementation. Yes, but you have some additional options that will produce real time shadows at virtually no cost.

That seems like a pretty reasonable price, if you have a finished game and are ready to publish.

Maybe. This is the kind of optimization you can't really predict ahead of time. You just have to start using the system, find the bottlenecks that arise, and then solve them. The best part of this is you get proper six-sided point light shadows, at a cost that is similar to hacks like the dual parobaloid technique, without the visual errors.

DXTC3 might be the same as DXTC5, with a 1-bit alpha channel. Not sure, though. Don't ever use DXTC1, the compression artifacts are awful.