You must be registered to see the links
You must be registered to see the links
As I have previously mentioned, I have been experimenting with interactive 3D scene rendering. The video shows what is currently possible to do using the system I have been developing. Basically, instead of just rendering a static image, it's now also possible to export the whole scene from Blender and render it interactively inside Ren'Py. In the video, you can also see very basic hair simulation and boob-physics. It also supports facial expressions (smile, frown, etc.), posing and animations. The visuals are not that great yet as the renderer I wrote is a pretty quick-and-dirty version with no fancy features. It can be improved or maybe even replaced with a proper, fully-featured renderer (maybe with virtual reality support).
There are plenty of nice things that 3D can provide, but also some disadvantages. I'll list some of them here.
Advantages:
-Interactivity is just cool. Moving things tend to be much more interesting than simple, static images.
-Different camera angles, poses and expressions are easy to do. No need to render separate images for all combinations.
-Pretty easy to do, it only takes a few clicks to export a typical scene and maybe ten minutes of extra work (for now).
-Animations can also be exported, so instead of using pre-rendered video files it's possible to add some interactivity, different point of views, free look etc. This could actually save some disk space as hi-quality videos can get pretty big.
-Interactive headpats and hair ruffling. And other grabby stuff.
Disadvantages:
-Interactive visual quality tends to be lower than in static images as the interactive rendering should be fast (preferably 30-60 fps). Static images can take hours to render and post-process so the quality tends to be a lot better. Still, the demo interactive renderer can be improved significantly.
-3D scenes require much more from CPU and GPU. Ren'Py is not really designed for this kind of stuff and the performance of the RenPy's scripting language (Python) is starting to limit things a bit. Static fallback images can be used by those who have slow/old computers.
-3D content takes much more disk space (estimated maybe 5-15 times more than static images). A typical 1280x720 static .jpg image takes around 0.2 megabytes and some of the scene backgrounds use additional metadata which pushes the image size to around 1.5 MB. In contrast, the 3D scene in the video takes about 11 MBs (compressed) but it also uses large 4K textures which can be made a lot smaller (although quality will suffer a bit). Also, some model geometry and textures can be re-used in multiple scenes which will also help.
-3D content used for interactive rendering sometimes requires a different license which tend to be much more expensive. An outfit might cost 5-20$, but the interactive licenses tend to be around 25-50$ and they add up quickly as each character needs multiple sets of clothing.
So, what do you think? Are you interested in seeing more interactive stuff? I might do a poll about this later to see what people think.
