HF2 3d model capabilities

Hi guys, been dead a long while working on small projects here and there...
I am picking up HF2 this week, and I am so looking forward to the 3d model support but I have a few questions for those already using it. The most important one is, how detailed can the object be?
I'm getting a friend who works on counterstrike maps to render them out as LWO's to use with hitfilm, my idea being that any maps made for video games are already the perfect indie virtual sets. I've toyed with it already using fraps to record camera moves in various source titles, but the trick has been trying to get the camera angles right when mixing with green screen footage, something I've always been stuck for space with. So i am really hoping that I can use the large levels in hitfilm so I needn't worry about getting the angle right in the green screen, as I can just adjust the 3d camera angle for the backdrop to suit.
Has anyone explored this option yet?


  • I haven't tried it out yet, but I think this would just depend on how much your machine is able to handle - enough RAM to work with a big model would probably be key.
  • HitFilm 2 will render a 3D model consisting of hundreds of thousands of polygons at interactive rates. Adding ambient occlusion, shadows, motion blur and depth of field will improve the 3D model's appearance greatly, but will impact performance. You can disable these options while editing to ensure that you can continue to work fast and just turn them on when you need to tweak the appearance of your 3D models.
    If very fast render times are important to you, then you can use only the minimum number of polygons in your 3D models to represent the detail you want. It is also a good idea to use a few small texture maps.
    If you are after high quality output then increase the texture map size and add more vertices to your model in the areas you would like more detail. HitFilm 2 also supports specular maps to add more detail to your models.
    If your entire scene does not render fast enough for you on your machine, then one approach you could take is to use a low polygon version of a large area for distant shots and improve the quality of the model (by increasing texture map size and polygon counts) for camera shots that are near to the 3D objects. How much your machine is capable of rendering interactively will depend on your GPU and RAM, as Robin suggested.
  • thanks for the replies. yeah they will be counter strike maps mostly, so no real need for extra detail. I'm sure if my machine starts to struggle I can get the models chopped into pieces and just use the rooms/sections that I need for the composite. I'm just glad that my idea of using them as virtual sets is actually viable :)

Sign in to comment