Image Processing Reference
There are conventions for mapping the cadence to the time code of the video. Your encod-
ing preprocessor that does an inverse telecine operation may look for this mapping in
order to do the best possible job. The frames are named A, B, C, and D. The A and C frames
are repeated twice; the B and D frames are repeated 3 times. The cadence is therefore
2-3-2-3, so the name “3-2 pulldown” is slightly misleading.
The clip may have been edited after the telecine conversion. This editing should be
done carefully to avoid interruption of the cadence. If the cadence is compromised, the de-
interlace code must offset the cadence detector from the beginning of the clip and attempt
to re-synchronize if necessary when there is an edit.
The solution is to perform the inverse telecine operation before importing the footage
into your nonlinear editing system (Premier, Avid, or Final Cut Pro) so that what you
are cutting is a sequence of progressively scanned video frames with no interlace artifacts
or pulldown cadence embedded in them.
Some encoding tools will make an attempt to detect this cadence slippage. Discreet
Cleaner, for example, is reasonably robust if it is not given a sequence where the cadence
is seriously damaged. It will cope with minor corruptions.
Some earlier versions of Adobe After Effects experience problems with this, and you
must check the processing of the footage and do some remedial work in order to get the
effect you want. While After Effects is not a compression tool, it is a very useful “Swiss
Army knife” for fixing things in video clips, so this may or may not be an issue that you
confront. If the cadence is correct and undamaged, After Effects will have no problems
processing continuous sequences of video several hours long. A cadence error in the mid-
dle will cause a problem. If this happens, fall back to Cleaner for that part of the process-
It's useful to have a variety of tools available to analyze the footage to inspect the
pulldown on NTSC content that may have originated at 24 fps.
Compromised Telecine Pulldown
Pulldown removal works best when it is processing telecine footage that has not been
processed any further once it has been converted to the interlaced format. You will have
problems where some telecine footage has been overlaid with a moving caption that was
rendered using interlaced motion because then you have a combination of both interlaced
and pulldown content within the same frame. Adaptive de-interlacing might be the only
way to process this properly.
Make sure you synchronize the sampling inputs on your video digitizer to the correct field
on the source. If this is configured incorrectly, the wrong field is drawn first. This causes a
nasty twittering artifact. There are two possibilities (Figure 33-10). How your video sys-
tem responds depends on the hardware implementation. Neither possibility does the