Inverse Telecine on Low Resolution NTSC Video

How Does Low Resolution Capture Affect Telecining?

Video that is captured at low resolution (XXX by 240) only contains one field.  Consequently, the video does not contain any interlacing artifacts.  Instead, one frame will be duplicated in each group of 5 frames.  Here are a couple diagrams that demonstrate why this is true:

The above is a video captured at full resolution (XXX by 480) with both fields.  The differences between each field in frame #2 and frame #3 lead to interlacing artifacts.  Now here is a diagram representing low resolution video:

As you can see, the bottom fields of the video are missing.  This happens because your capture card cuts out one of the fields when you capture at low resolution.  If we treat the top field as the entire frame (which it now is), you can see that frame #1 is repeated in frame #2.  This frame duplication occurs once in every group of 5 frames.

How Do Most Programs That Inverse Telecine Deal with Low Resolution Video?

VirtualDub and most other most other inverse telecine programs treat low resolution video as if it had 2 fields.  While this is not true, it allows low resolution video to have its telecining removed in the same manner as full resolution video.  Here is how most inverse telecine programs see video captured at low resolution:

Source video

Basically, all the odd horizontal lines in the video are treated like a top field and all the even horizontal lines are treated like a bottom field.  When the program correctly inverse telecines a 240 line video, it discards the top "field" in frame #1 and the bottom "field" in frame #2.  It then combines the remaining bottom "field" in frame #1 and top "field" in frame #2, producing a video that looks like this:

By eliminating two fields, one frame is removed from the source video.  The result is an output of 4 frames that correspond directly to the original film frames.  This is the desired output of inverse telecining.

How Does RealProducer Inverse Telecine Low Resolution Video?

As far as I can tell, RealProducer uses a novel approach when inverse telecining low resolution video.  Instead of splitting the video into 2 virtual fields, it actually views the video as consisting of whole frames.  It then analyzes the frames to find out which ones are duplicated.  When it finds a duplicated frame, it simply removes the frame, as you'd expect.  However, the interesting part is what happens when RealProducer can't find a duplicated frame.  If RealProducer cannot determine which frames are duplicated, it doesn't cut out any frames.  As a result of not removing frames in certain areas, the frame rate of the video is ~30 fps in some areas while it is ~24 fps in others.  RealProducer can do this because RealMedia files can have a variable frame rate.  AVI files, on the other hand, are fixed to a certain frame rate.  If an inverse telecine program for an AVI file didn't remove exactly 1 of every 5 frames, it would lead to audio/video desynchronization.

Should I Inverse Telecine Low Resolution Video?

Since low resolution video only contains one field, there are no interlacing artifacts.  Therefore, the advantages to inverse telecining a low resolution video are a reduction in file size and a video that plays more smoothly.  However, there can be one disadvantage to inverse telecining low resolution video.  If you are inverse telecining an AVI file and the inverse telecine program eliminates the wrong "fields", interlacing artifacts will actually be introduced to your video.  (I explain why this occurs in the section: "What Happens When a Program Incorrectly Inverse Telecines a Low Resolution Video?".)

If you use RealProducer to inverse telecine your video while compressing it to a RealMedia file, then you should definitely inverse telecine your video.  Thanks to RealProducer's unique way of inverse telecining a video, interlacing artifacts are never introduced.  Therefore, the file size will be reduced and the video will play more smoothly, and you don't have to worry about interlacing artifacts.

If you are using a program such as VirtualDub to inverse telecine an AVI file, the answer is not as simple.  The advantages are that your video will have a smaller file size and will play more smoothly.  The disadvantage is that interlacing artifacts will appear in places where the inverse teleciner messed up.  Unfortunately, trying to deinterlace your video to remove the artifacts is bad idea.  Since your video is already low resolution and because deinterlacing removes up to 50% of the detail in a video, I don't recommend deinterlacing low resolution video.

However, I think that the advantages of inverse telecining a low resolution AVI video outweigh the disadvantages.  The inverse teleciner doesn't mess up too often, so few frames will have interlacing artifacts introduced.  Also, the interlacing artifacts usually aren't that visible unless you are specifically looking for them.  Therefore, I do suggest that you inverse telecine your low resolution video.

How Do I Inverse Telecine Low Resolution Video?

If you are using VirtualDub to inverse telecine a low resolution video, the process is identical to the method used to inverse telecine full resolution video.  See the section called "How Do I Inverse Telecine Full Resolution Video?" for instructions on how to do this.

If you are using RealProducer to inverse telecine your video, you simply need to enable the Inverse-Telecine filter.  You can do this in RealProducer by clicking on Options -> Preferences -> Video Filter, checking the box labeled "Inverse-Telecine", and hitting OK to exit the Preferences screen.  Then you just encode your video as you normally would.