I'm working on a project with very limited texture space for a large object. I can have one 1024 map and one 512. One needs to be diffuse and the other for normals. In the Infinity Blade talk at GDC, they said that they used large normal maps and small diffuse maps for their textures to maximize the detail for a small screen. I would have thought to do it the other way with a large diffuse and smaller normal map. So my question is, which way is considered standard for a situation like this? I plan to try both anyway, but I am curious as to how others have done it in the past.

First thing that comes to mind for solving this sort of thing is using modularity of some sort.

As for the reasoning behind using bigger normal maps than diffuse could be that the surfaces become more dynamic in different lighting conditions, in turn building on top of existing diffuse detail.

Its really hard to say and might be worth it to test a few things out, may be more a stylistic choice rather than only 1 answer.

Another thing you can do to maximize it I think can rely on the shader network.

If you use vertex colors combined with greyscale texture maps you could potentially fit 3 texture sheets work of details into 1 RGB texture map. Just make sure the shader desaturates the texture and lets you pick a color channel to use. At least I think that could work. But that also relies on what type of environments and stuff your making.

That really isn't that tight of a budget. :P

The question of which for what comes down to what you are making. If you have what are essentially large blocks of similar colors, or smooth gradations, you can get away with smaller diffuse maps and letting the bilinear filtering work for you, while getting a nicely detailed surface with a larger normal map.

If you have largely flat or smoothly changing surfaces with limited small scale detail, then you can use the higher amount of color information and go with the smaller normal map.

Pfft. Back in my day "tight" was 32*32 pixels.