Inverse Telecine (a.k.a. 3:2 Pulldown Removal) on Full Resolution NTSC Video



There are several disadvantages to NTSC video that has been telecined.  Specifically, it produces interlacing artifacts on a computer monitor, makes a video play less smoothly, and increases the file size.  Fortunately, it is possible to remove the duplicated fields that were created during the telecining process and reconstruct the original film frames.  The process of undoing the results of telecining is known as inverse telecine.

The telecine process uses both fields of a video to alter the frame rate.  Therefore, the best way to remove telecining is by capturing your video in full resolution (XXX by 480).  When you capture a video with a vertical resolution of 480 lines, you are capturing both fields of the video.  Having both fields of the video present during the inverse telecine process increases the accuracy.

How Do Programs That Inverse Telecine Deal with Full Resolution Video?

To inverse telecine a full resolution video, a video editing program must detect the two duplicated fields, remove them, and combine the remaining fields into a single frame.  Here are some diagrams that show how this is done:

Frame #1
Frame #2
Frame #3
Frame #4
Frame #5
1T 1T 2T 3T 4T
1B 2B 3B 3B 4B

Source Video

As I mentioned earlier, field 1T in TV Frame #2 and field 3B in TV Frame #3 are duplicates.  When a video editing program correctly detects that these fields are the duplicated ones, it eliminates them.  Then it combines the remaining bottom field of TV Frame #2 (2B) with the remaining top field of TV Frame #3 (2T) to produce an output that looks like this:

Frame A Frame B Frame C Frame D
1T 2T 3T 4T
1B 2B 3B 4B

You can see that the duplicated fields have been removed.  Because of this, there are only 4 output frames, and they correspond directly to the original film frames.  This is the goal of inverse telecining.

How Do I Inverse Telecine Full Resolution Video?

Removing the telecining of a video is actually quite easy, thanks to advanced video editors.  In my opinion, the best program to inverse telecine a video is VirtualDub.  Here's how you do it:

1)  Open your video with VirtualDub.

2)  Click on Video -> Frame Rate to bring up the Video frame rate control dialog box.

3)  Go to the section labeled "Inverse telecine (3:2 pulldown removal)", select the option "Reconstruct from fields - adaptive", and hit OK.  This option automatically detects the telecining pattern and adjusts to any changes.  While it is also possible to use the "Reconstruct from fields - manual" option, this method usually won't work.  This is because the telecining pattern often shifts during a video, which causes the manual method to produce incorrect, interlaced frames.  I explain why these shifts occur in the section entitled "Why Do Programs Mess Up When Doing an Automatic Inverse Telecine?".

4)  Finally, apply any video filters that you wish to use, set your desired video and audio compression, and save the AVI.  The AVI will be inverse telecined as VirtualDub processes/saves the video.