How IVTC Restores 24 fps Film from 29.97 fps VOB

Inverse telecine (IVTC) is the reverse-engineering process used to reconstruct original 24 frames-per-second (fps) progressive film from a 29.97 fps interlaced NTSC video, such as a DVD VOB file. When motion pictures are transferred to NTSC video, a technique called 3:2 pulldown stretches 24 film frames across 60 interlaced fields (29.97 frames) per second, introducing duplicate fields and interlacing artifacts. IVTC detects this repeating cadence, pairs the original complementary fields back together into complete progressive frames, and discards the redundant data to restore pristine 23.976 fps video.

The 3:2 Pulldown Mechanism

To understand how IVTC reverses the process, consider how film is converted to NTSC video:

1. Four original film frames (labeled A, B, C, and D) must occupy the duration of five television frames (ten interlaced fields).
2. Each film frame is split into top and bottom fields.
3. The fields are distributed in an alternating 3-field and 2-field pattern:
   • Frame A contributes 2 fields (Top A, Bottom A)
   • Frame B contributes 3 fields (Top B, Bottom B, Top B)
   • Frame C contributes 2 fields (Bottom C, Top C)
   • Frame D contributes 3 fields (Bottom D, Top D, Bottom D)

This creates five interlaced video frames from four film frames. While frames composed entirely of "A" or "D" fields remain visually progressive, the frames containing mixed fields (like Bottom B paired with Top C) exhibit heavy comb-like interlacing artifacts during motion.

Soft Telecine vs. Hard Telecine in VOBs

VOB (MPEG-2) files handle telecine in two distinct ways:

• Soft Telecine: The video is physically encoded on the disc as native 23.976 fps progressive frames. The MPEG-2 bitstream includes metadata flags—specifically repeat_first_field (RFF) and top_field_first (TFF)—that instruct the DVD player to generate the 3:2 pulldown in real time during playback. For soft-telecined VOBs, IVTC does not require complex video processing; the software simply ignores the repeat flags and extracts the native progressive frames directly.
• Hard Telecine: The 3:2 pulldown was burned directly into the video stream before MPEG-2 encoding. The VOB contains an actual 29.97 fps interlaced stream. Restoring these files requires true algorithmic IVTC.

The Stages of Algorithmic IVTC

For hard-telecined footage, the IVTC algorithm operates through three precise steps:

1. Cadence Detection

The algorithm analyzes consecutive fields across the video stream to determine the 3:2 pulldown pattern. It looks for matching motion vectors and pixel differences to locate the duplicated fields that occur every five frames. Because film edits, commercial breaks, or hybrid video sources can cause cadence breaks, modern IVTC filters continually analyze the stream rather than assuming a static rhythm.

2. Field Matching (Weaving)

Once the pattern is locked, the filter groups complementary fields. It extracts the top and bottom fields that originated from the exact same exposure of film and weaves them together:

• Fields derived from Frame A are combined into a progressive Frame A.
• The matching fields for Frame B are combined into Frame B.
• The split fields for Frame C and Frame D are properly aligned and combined.

At this point, the interlacing artifacts disappear because no reconstructed frame contains fields from two different moments in time.

3. Decimation

Combining the ten fields yields five reconstructed frames, but one of these frames is an exact duplicate caused by the extra fields generated during the "3" phases of the pulldown. The decimation step identifies and drops this fifth redundant frame from every group.

By removing one duplicate frame out of every five, the frame rate is mathematically reduced by 20%: from 29.97 fps down to the original film standard of 23.976 fps (often rounded to 24 fps). The final output matches the motion, cadence, and progressive structure of the original theatrical film.