7 Replies Latest reply on Jan 18, 2009 5:36 AM by (Steve_Patterson)


    (Steve_Patterson) Level 1
      Ok, have you guys encountered this frame rate before?
      I was handed a piece of footage (quicktime) that appears to have that framerate encoded in the metadata, and is a 1920x1080 file.
      It "appears" to be progressive, despite it's non-integer frame rate.
      I've tried dropping it in a comp native to it's interpretation, and I'm seeing stutters that look to be transfer/frame rate/field oriented.
      I've tried a slew of different interpretations, field separations, etc... to no avail. I've tried searching community help and the interweb for 24.975, and haven't gotten any solid answers. When trying to interpret the footage with fields (upper and lower), I've tried doubling my comp framerate to see if the field-ordering had any effect, and it did not appear to.

      It seems to me that it would be an interlaced PAL framerate... Can you guys give me any suggestions about how to fix (properly interpret) the stutter? Especially our friends in PAL land, or those NTSC folks with more PAL knowhow than myself?

      It could always be a case of "cr4p in cr4p out", but I want to make sure that's the case before telling the client that.

        • 1. Re: 24.975
          Level 1
          The test to determine whether or not the footage is interlaced is easy. You only have to open up the footage interpretation window and select separate upper, then drop the footage in the new comp icon at the bottom of the Project panel, double the frame rate of that comp, then step through the footage a frame at a time using the page down key. If you're getting double frames or it seems like you're skipping frames once in a while it could be that the footage has a 3:2 pulldown scheme.

          If the footage is progressive you'll see a series of identical frames. If the footage is interlaced each frame will be unique. If the motion goes back and forth then the footage is interlaced lower first. If the motion is only forward then the interpreted correctly. This is easiest to see if there is significant motion in the frames. If you get two identical then one with movement and then two identical every once in a while, then your footage has 3:2 pulldown and you'll need to figure out the cadence.

          Determining the proper frame rate is a bit more difficult. In fact, without some kind of real time clock on the footage, ie,. encoded time code, you can't check the time against real time. The only difference you'll notice is a slight pitch change in the audio. As long as you properly interpret the footage so that you get only complete frames, make sure the comp frame rate is the same as the footage, and check the maintain frame rate of nested comps option you're AE projects will not have any missing frames. It may playback a little too fast or two slow, but it won't have any missing frames.
          • 2. Re: 24.975
            David Wigforss-Hv1BNN Level 2
            Funny, for this freelance gig, I have the same. 1920x1080 24.975 1.0par. I was thinking that they gave me the incorrect specs, but figured I'd do a test and see if they can import it back. Mine looks low quality/compressed, even though they gave me lossless qt:anim. I guess mine originated from RED, so I'm not sure how they went from RED -> QT:Anim (1920x1080).
            • 3. Re: 24.975
              (Steve_Patterson) Level 1
              Thanks guys!

              @ Rick, when dropping it (the UF separated footage) in a comp & doubling the frame rate, I'm seeing one frame with motion, then a static frame, then the next frame with motion, then static, etc...). I did the test in a section with significant camera motion. This leads me to believe that the quicktime is progressive, as I thought originally. However, it still seems, despite the behavior I'm seeing as I step through the double-fps comp, that there are frames missing. For example, the camera pan should be a relatively constant speed, but, it would appear that every now and then there are frames missing because the move will "skip" ahead faster than it should. If I were to remove pulldown, the resultant effective frame rate would be 19.980. I've tried every possible pulldown-scheme, and no dice. It just looks worse.

              I'm at a loss. I'm guessing that they had already somehow converted the footage between shooting it and giving it to me, and in the process have dropped some frames.

              @ David, are you seeing similar stuttering?
              • 4. Re: 24.975
                David Wigforss-Hv1BNN Level 2
                Away from the project at the moment, so I'm not sure how many frames are inbetween. But yes, there are frames every so often that are held for 2 frames. Maybe 5 frames and then a held frame, etc.
                • 5. Re: 24.975
                  (Steve_Patterson) Level 1
                  Thanks again, David. Yours actually sounds like one of the pulldown schemes (WWWWS). I'd be interested to know how it turns out.
                  • 6. Re: 24.975
                    David Wigforss-Hv1BNN Level 2
                    I tried doing a render at lunchtime, and somthing's jacked! My footage is interpolated as 24.975. My comp is 24.975. In renderqueue, best settings said comp rate was 24.98. I changed that to 24.975, and still my layer isn't syncing to the video layer.

                    It looks like problems I've experienced before where the footage is 23.976 but I created my comp at 24. But I've made sure everything was the same for this project (24.975fps).

                    Trying to get together with client to figure out what's going on (IE why is my footage 24.975 as opposed to 29.97 or 23.976).
                    • 7. Re: 24.975
                      (Steve_Patterson) Level 1
                      I hear ya', David. At this point, I've tried smoothing it with Timewarp, and still no dice. I'd be interested to hear what your client says about the source. Thanks for posting back!