Understanding 24p 2-3 pulldown
When a video is shot in 24 frames per second (fps), in order to convert it to normal NTSC frame rate, the video is converted using a 3:2 pulldown.
A frame of NTSC video consists of 2 fields interlaced. These two fields add up to one frame, but are displayed 1/60 second apart. The odd lines of the frame are displayed first and then the even lines are displayed. The human eye merges these two fields into a single frame. However, if you pause the video/DVD, you can often see interlace artifacts in areas where something is moving. These artifacts look like a comb, where every other line is displaced from the lines around it.
What this means is that every odd frame of the 24 fps is copied to 3 fields in the video, and then every even frame is copied to the next 2 fields. In a best-case scenario, what you will see is 3 normal frames and then the fourth frame will be duplicated. in a worse case scenario, only every fourth frame will be clear, the rest will be some combination of two different frames.
NTSC Frame | Non-Blend translation | Frame Blend Translation |
---|---|---|
0 | 0 | 0/0 |
1 | 0 | 0/1 |
2 | 1 | 1/2 |
3 | 2 | 2/2 |
4 | 3 | 3/3 |
5 | 4 | 4/4 |
6 | 4 | 4/5 |
7 | 5 | 5/6 |
8 | 6 | 6/6 |
9 | 7 | 7/7 |
10 | 8 | 8/8 |
11 | 8 | 8/9 |
As you can see, the non-frame causes a stutter every four frames (or eight times a second) and the frame-blending method causes two frames to be blurry followed by three frames not blurry.