(Page 1 of 2 pages for this article  1 2 >)

Filed under: Editing •Motion Graphics •Post Production •

Compression Artifacts & Pulldown

This started as a quick post about how to gain finer control over the compression settings in the QuickTime dialog. But before we can get there, we first need to talk talk about how 3:2 pulldown works. (Trust me; it all ties together; it was also a good little mystery.)

I recently gave a training session at a local studio, and at the end they were invited to trot out their Barney Stumpers (questions about why something went wrong, how something works, etc.). For one stumper, a user had some footage with 3:2 pulldown, and after pulldown was removed, he noticed that an after-image of the previous frame appeared in the next frame after an edit. Why?

Pulldown and Split Frames

3:2 pulldown is the process of spreading 4 frames of film across 5 frames of video to help resolve the frame rate differences between the two. The most common form spreads 4 whole film frames across 10 video fields, following a 3/2 alternating pattern: the first film frame is spread across 2 video fields, the next film frame is spread across 3 video fields, and so on. Here is a diagram of how the process works:

Sometimes, the resulting video frame is “whole” - both of its fields came from the same original film frame. Sometimes, the resulting video frame is “split” - one field came from film frame, and the other field came from the next film frame. But as long as you know the pattern, you can reverse the process and get back to the original whole film frames.

A potential problem arises if you take footage with 3:2 pulldown and compress it using a codec that is unaware of the interlaced nature of video - one that instead compresses each whole video frame as if it was a single image from the same point in time. Compress it too hard, and compression artifacts can smear bits of one field into pixels of the adjacent field. This is a problem if you’re compressing a “split” frame as described above, as it contains two different images. When you later remove the 3:2 pulldown sequence, suddenly you find bits of one film frame appearing in an adjacent film frame. This is particularly obvious is the scene changed between those frames - and that’s what the user was seeing:

The two frames above show scenes before and after an edit. After applying 3:2 pulldown, compressing the heck out of it, and removing pulldown, there is a chance (depending on where the cut lands) that the result will be a frame that looks like the one at left, which includes a ghost of one frame superimposed over the prior frame.

In this case, the user had chosen the PhotoJPEG codec, which was designed for whole images - not interlaced footage. Despite this shortcoming, it’s a codec we personally like, and some video card manufacturers such as Blackmagic also use it. But if you reduce it’s quality setting too much, you’ll get the smearing described above.

When the user asked me how to cure this problem, one solution I suggested was setting the quality higher. And here’s the punch line: He didn’t know you could change the quality settings of codecs such as this. Which brings us to the real point of this post - which is continued on the next page (so much for this being a quick post).

(Page 1 of 2 pages for this article  1 2 >)

                   

 

AJA and Sound Devices embrace Sony NXCAM’s timecode-over-HDMI Allan Tépper | 02/29 Free firmware updates enable timecode-over-HDMI from NXCAM, but is that enough? Many ProVideo Coalition readers may recall my article called Untapped features in Sony NXCAM’s new HDMI output from June 2011. At that point, I surveyed several…

PsF’s missing workflow, Part 6: Tépper asks the camera manufacturers… Allan Tépper | 11/14 An open letter to professional AVCHD camera manufacturers In parts 1-3 of the PsF’s missing workflow series, we introduced the terms benign PsF & malignant PsF, and revealed the PsF status…

PsF’s missing workflow, Part 5: Átomos hires a stripper! Allan Tépper | 11/11 Átomos hires a stripper to counteract malignant PsF! In part 1 of PsF’s missing workflow, we introduced the new terms benign PsF and malignant PsF (Progressive Segmented Frame), reviewed their…

You must be registered to comment. This is an effort to reduce spam. Please REGISTER HERE.

Name:

Email:

Location:

URL:

Smileys

Remember my personal information

Notify me of follow-up comments?

Submit the word you see below: