Announcement

Collapse
No announcement yet.

FilmRender: "film look" and configurable deinterlacing!!

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    My question is, what exactly is a "film look"?
    From my understanding, the reason video looks different from film is a) Film is 24 frames/sec vs. 30 frames NTSC which does not create hyper-smooth movement that you see in video
    b) the colour range: film is able to capture a huge spectrum of colours as it is an actual photo created from light whereas video is an interpretation of light and colour through electronic sensors

    Perhaps the interlacing or rather lack of interlacing is the issue too...
    but other than that, these are not things that can be changed with a computer program. The colour depth is limited to the source and to achieve the correct framerate would mean throwing away every 5th frame.

    Am I correct or waaay off?

    Comment


    • #17
      Jeff - let me know how you get on. Can you tell me where to pick up the adaptec DV codec, I reckon there are some other people who would find this usefull.

      intomisery -
      a) correct
      b) correct, kind of. Yes film can do this, but it suffers the same limitations as video, because it is telecined and played back on a TV.

      Now, it's important to understand what FR is doing. Firstly, it is NOT doing and inverse telecine (3:2), although that can be done, programs like After Effects and DigiEffects cinemotion do that. The reason I haven't done it in FR is because I live in PAL land. I would never be able to test it! When/if FR makes some money I can buy some NTSC equipment to work on it.

      FR de-interlaces the fields. See http://www.nurkware.com/filmrender/d...nterlacing.htm
      this gives you 25fps on PAL, 30fps NTSC.

      As for colour changes, it is more a change of colour and brightness response. Most video makers cannot afford big lighting rigs and a lot of time controlling the light (which is more difficult than film in the first place - video has a much lower contrast ration). To produce those deep blacks, why not just adjust the brightness at the lower end of the spectrum? And the subtle colouring on film stock makes for more interesting viewing - take say 'Taxi' (the film or) 'The thin red line' or 'saving private ryan'. No linear light responses there!

      Regards
      Ned

      Comment


      • #18
        so in order to achieve a better film look, I should bring in my video at say 24 fps, then use an alternate program to do the 3:2 pulldown?
        what about the sharpness of the image.. I have not looked in detail at your program, but by changing the blur, letting the sharp, harsh lines slightly blend into each other, I should be able to get a filmy look?

        I'll have to do a bit more research and read your manual, but I'm still unsure about all this...

        thanks though

        Comment


        • #19
          Hi there, sorry for the delay in replying, been a busy week.

          I don't know how you could get your video at 24fps, but if you could that would be great.

          When it comes to the 3:2 pulldown, I'm no expert as I live in PAL land and so don't get a chance to experiment. Try just using the smart blur option and see how you find that. Maybe you might want to feed the output of FR into After Effects to add the 3:2 pulldown, maybe not.

          The real thing about apps like FR is that they make video look like 'not video'. To 99% of people that means film and although a trained eye can spot the difference easily, most of your audience will probably not be able to notice the difference.

          >what about the sharpness of the image..

          The 'smart blur' options are only there to blend the interlaced fields together to give you de-interlaced video.

          No matter how you process your interlaced video to de-interlaced video, you will experience some kind of trade off. In many ways, we are trying to conjure something out of nothing. Plain de-interlacing (i.e. middle field is the average of the field above and below) retains sharpness, but becomes very jaggy and looses detail.

          Smart blur merges the fields together (every field is the 50% original, 25% above, 25% below, if merged) and can retain more detail than de-interlacing, without jaggies. But can be a bit 'smooth' (although I have implemented a configurable vertical sharpen as well to offset this)

          Regards
          Ned

          Comment

          Working...
          X