This is relevant to my interests.
First, I must applaud the effort and pace of work.
Then to business: the MediaVP's offer a wide range of very, very good stars to be used. They are basically a result of
this thread and they work very well in-game; I would love to see how they look in cutscenes.
There are also better starfields than what you have been using. The background stars seem oddly uniformly bright and they have thus strange blurred thing around them, while in space they would be crisp and come in multiple brightnesses and the brightest stars would usually also have a discernible hue to them, either reddish or bluish (yellow stars appear white to human eyes, and dim stars are usually too dim to register as colours in our eyes due to the relatively high threshold of colour detecting cells).
Here's my view on a generic starfield; something a lot like this is used by the MediaVP's. However the version in MediaVP's turned out to divide opinions quite strongly, so I'll just say that this or something similar is what I would prefer to see:
(not using lvlshot for that one because it would basically just show a black square due to massive downscaling...)
This is a 4096^2 resolution starfield and can be mapped cubically on a sphere to effective 16384 pixels of horizontal resolution. That means you can use field of view as low as 22.5 degrees and still take 1024^2 renders that maintain singular stars as pixels rather than upscaled dots.
Cropping a 2048^2 section of this map would work just as well down to 45 degrees of field of view angle. The resolution you would choose should depend on the view angle, because starfields are things that typically don't downscale very well at all, so mip mapping is somewhat problematic.
Nebulas, hmm. Unless you are intentionally emulating the look and feel of FS1 nebulas, I'd say having a bit of additional detail couldn't hurt. However, as I've come to notice, convincing nebulas are damn hard to make. The best results I've had are using GIMP's Plasma render, then recolouring it to certain hues. I really have not devised a method yet to mimic the complex, real nebular gas formations yet, though.