Quote:
Originally Posted by Bob Kittleson
(Post 873244)
Aaron, I mean no disrespect but you haven't provided any sources to back up your assertion that fixed pixel displays produce 30fps output from 60i input streams. I can't prove you incorrect either, but intuitively what you said doesn't seem to add up. How can the following points be explained?
|
http://www.hqv.com/technology/index1...TOKEN=16476801
This is a good read on the subject of de-interlacing by HQV, one of the leading vendors of video processing technology. As you say, you don't believe half of the vert rez is missing from interlaced video while displayed on your progressive set, which means perhaps this blurb is a bit dated, since no one seems to be using non-motion adaptive de-interlacing any longer in these displays or BD players.
Quote:
- If displays process 60i as you describe, then half of the motion resolution would just be thrown away for video content that was actually recorded in 60i. I don't know about anyone else, but when I replaced my CRT display with a fixed pixel display, I did not notice any loss of motion smoothness. I would think that a reduction by half would be noticeable.
|
On the contrary. Motion resolution compared to what? Interlaced CRTs? I don't like the term applied to fixed pixel displays because you have ONE resolution - the native rez of the screen. And that's all you've got. Proper de-interlacing attempts to minimize motion artifacts that are created when the two fields are recombined to display as a full frame on a progressive display, yielding back the original 30 fps. Again, read the HQV blurb...
Quote:
- The latest camcorders continue to record in 60i as their default mode, even if capable of 30p recording. Canon does not market 30p as being equivalent or better than 60i when played back on a "good" display. We know that recording in 30p eliminates deinterlacing artifacts, so why would anyone want to record in 60i unless it provided some benefit in terms of motion resolution? You could argue that this is because there are so many displays that don't handle 30p 2:2 properly, but still I would think that if ANY displays existed on which 30p was flat out better than 60i, then Canon would market that fact.
|
30P is flat out better in all things except compatibility when compared to 60i and viewing on progressive displays. This is factual and not even debatable. Due to the compatibility issues that I and many other professional reviewers have raised, manufacturers are clearly sticking with outdated technology rooted in NTSC. Seriously, how can anyone argue that capturing half of a frame and then capturing the second half of the same image 1/60 sec later is better than simply capturing the entire frame all at the same time? If you capture progressively, you don't have to deal with this de-interlacing mess and trying to minimize the problems you get when recombining the fields. It all just goes away, LOL.
Quote:
- Our buddy Ken Ross is an experienced video enthusiast who has tested both raw 30p and 60i video footage on a Kuro, which is one of the best displays available. He has clearly said that 60i video looks smoother to him.
|
I would bet money that his Kuro cannot properly de-interlace 30P in 60i video. And by properly, I mean (1) detect lack of interframe movement within the two fields comprising that frame and necessarily employ a weave de-interlace process, and (2) frame double the original 30 fps to match 60 cycles per sec of the display.
Quote:
- The article below notes that some crummy displays simply interpolate each individual field to form a full frame. This results in poor spatial resolution, but the full motion resolution of the 60i input stream is preserved. If crummy displays can retain full motion resolution, I find it hard to believe that good displays would not do likewise. I have not seen any articles that say motion resolution has to be sacrificed to get full spatial resolution.
http://www.hometheatermag.com/hookmeup/1107hook2/
|
IMO, this occurs when the video processor "gives up" on properly de-interlacing interlaced fields and is certainly NOT desireable. If that is "success" in your book, then you may as well stick with 720/60P and get real HD video.
Quote:
- This is pure speculation on my part, but if a good display can properly deinterlace field pairs within a frame, wouldn't it be possible to likewise deinterlace fields from adjacent frames?
For example:
Input frame 1 field A + input frame 1 field B = Output frame 1
Input frame 1 field B + input frame 2 field A = Output frame 2
Input frame 2 field A + input frame 2 field B = Output frame 3
Input frame 2 field B + input frame 3 field A = Output frame 4
...
This approach would in theory yield 60 unique output frames per sec on the display. It doesn't seem fundamentally different or more difficult than only deinterlacing intra-frame field pairs, other than requiring more processing power because more deinterlacing is being done.
|
Again, then you're technically not trying to combine fields comprising the original frame as captured by the video camera. You're trying to fabricate a frame that did not exist ala 3:2 pulldown (the problems of which have been identified for years and thus the whole reason for removing pulldown while editing 24P video, rendering out to true 24P BD, and then buying electronics that will pass ONLY 24 fps to the display at which point the display then frame doubles, triples, quadruples, or quintuples that original video or film frame in order to provide a better viewing experience.