Full HD and De-interlacing, Part 2
In the last issue, I discussed Full HD, a marketing program that promotes displays and camcorders, including consumer models, with 1920x1080 60p display/capture capabilities. I also started to describe de-interlacing, and in this edition I will describe several different methods of de-interlacing and resume the Full HD discussion.
Bob De-interlacing
Here is the correct description of bob de-interlacing, which I had incorrectly described in an earlier article. As each 540-line field is input to the de-interlacer, an interpolator generates the needed lines from pixels in the lines directly above and below new lines.
xmlns:w="urn:schemas-microsoft-com:office:word"xmlns="http://www.w3.org/TR/REC-html40">
|
|
|
|
|
l |
|
|
|
|
|
|
||
l |
|
l |
|
|
|
l |
|
|
|
|
|
||
l |
|
|
|
l |
|
l |
|
|
|
|
|
||
l |
|
|
|
|
|
l |
The result is a synthetic 1080-line frame that is sent to the panel each field-time — every 1/60 second. To some, the constructed image seems to vibrate up and down, hence the term “bob.” Diagonals are not handled well because the interpolator cannot create pixels, as there are no pixels above and below the locations (yellow cells) where they should be placed. Effective vertical resolution is 540 TV lines per field-time.
2D FIR
As each 540-line field is input to the de-interlacer, it is vertically scaled to a 1080-line frame using a 2D FIR filter.
xmlns:w="urn:schemas-microsoft-com:office:word"xmlns="http://www.w3.org/TR/REC-html40">
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The generated pixels (red dots) yield a frame with relatively few jaggies. Each upscaled 1080-line frame is sent to the panel each field-time — every 1/60 second. Vertical resolution will be approximately 750 TV lines per field-time.
Frame-Based Motion Adaptive De-interlacing
A frame-adaptive de-interlacer uses logic to measure motion between fields. For static frames — weave is employed. For dynamic frames, either 2D FIR or a bob is used.
Region-Based Motion Adaptive De-interlacing
A region-adaptive de-interlacer uses logic to measure motion within regions of a frame. Weave is used for static regions; for dynamic regions, either 2D FIR or bob is used. The advantage of this system is that only areas with motion lose vertical resolution.
Motion Vector Adaptive De-interlacing
A Vector Adaptive interpolator uses memory to hold four fields. Logic measures motion between fields. For static video, weave is employed. For dynamic video, samples come from the current plus a previous and/or a future field. Dynamic video will have approximately 960 TV lines of vertical resolution per field-time.
Assuming a display has a 1920x1080 panel and an optimum de-interlacer, are there other “Full HD” issues to consider? Absolutely. There remains the open question of what type of video connection — 1080/60i, 1080/60p or 1080/24p — is required to obtain the maximum quality from a Full HD display.
The question of what type of video can be input by a display has no relevance for 1080/60i video; a standard 1080i connection is all that''s required. The question is, however, relevant to 24fps film or video. Both have an inherent artifact from their very low temporal sampling rate. The artifact is called “judder.” (Some believe judder is a key part of the look of film. Others point out that most cinematographers take great care to prevent judder — and so hotly dispute this contention.)
Judder is also the name given to the “look” of 24fps media after 2:3 pulldown has been applied to convert it to 60i. In this process, two “judder” (red) frames are created for each five video fames.
xmlns:w="urn:schemas-microsoft-com:office:word"xmlns="http://www.w3.org/TR/REC-html40">
Middle Ad
CPN Newsletter
FREE eNewsletter
Subscribe to our FREE eNewsletter
Subscribe to our FREE eNewsletter
Middle1 Ad
print edition
