Actually, I mean film is cheaper because under highly complicated circumstances, the standard workflow is less risky. I'm not saying that it's cheaper in princi

Actually, I mean film is cheaper because under highly complicated circumstances, the standard workflow is less risky. I'm not saying that it's cheaper in principle, or that there for a given situation there isn't a specific solution that's better; But when workflow is everything, it definitely is. It's much cheaper having to spend an hour on lighting, not worry about motion control and ask for 3 different setups in the telecine pull list, to have enough dynamic range to fit everything into the shot. With RED, even though results are very good (they are, however, short of amazing: it's terribly noisy, but then again so is underexposed film), the dynamic range is lower (and the dynamic density is uniform, which isn't that good), so keying becomes an issue, and so do CCD artefacts - even more so when renting 422 equipment. The complexity/risk isn't necessarily prohibitive, but definitely less lucrative. That's why I say 'cheaper', not 'better'. On the compositing side, software has been great for quite some time. I'm aware of only a handful of truly novel keying algorithms published in the past 6 years, and none that can deal with digital artefacts very well (if you know of any, please tell me. Better yet, tell someone who works for The Foundry or Autodesk. Much appreciated. x), that's why so many commercials are still shot on 35mm instead of digital.