Originally Posted by Joe Bloggs
Why aren't the BDA adding a 5120x2160 (or other wider than 3840) square pixel mode for the "4K" Blu-ray spec? Wouldn't that potentially give better picture quality than storing it in the 1.78:1 frame?
Do you have any demo clips that we can view of video encoded in the MFE process? And is it absolutely certain that the black bar areas won't have any visible difference to non-MFE encoded video when played on non-MFE capable players/displays and that there will be no visible degradation of the picture on non-MFE players/displays compared to non-MFE encoded video?
Assuming the BDA added anamorphic modes to the Blu-ray HD standards (for 1080p and 2160p) so that it stored the video in 1920x1080 or 3840x2160 but with a flag/flags to say they were non-square pixels, wouldn't encoding that anamorphic video be less efficient than normal 1.78:1 square pixel video (without black bars)? eg. wouldn't things like rotations (eg. for a 2D wheel) be harder to predict if it was stored anamorphically instead of with square pixels?
For the record, we are only discussing square pixels. MFE works with square pixels going in and going out. The end result after decode is a true anamorphic image at 1920 x 1080 using square pixels.
As to what the BDA is doing, I can't say definitively, but I would guess that they are not really looking at 5120 x 2160 because there would be very little application for it. Unless 21:9 displays start taking off in a big way, the reality is that just about any display in the future is going to be 16:9. That plus the fact that human beings perceive detail much more acutely in the vertical than in the horizontal. An anamorphic 16:9 image - even one at 1080P - will reveal almost all of the detail inherent in the source material, particularly if the source material is 35mm film.
One of the things that is being discovered very quickly is that perceptible
visual detail from 35mm film scans maxes out somewhere around 1080P, perhaps a bit higher, but nowhere near 4K levels. From the testing being done in Hollywood (some of which I have witnessed), 4K resolution seems to be considerable overkill when it comes to 35mm film. Once you reach a certain threshold, limiting factors such as camera judder, MTF and film grain just obliterate the fine detail you would want in a 4K transfer (70mm is a different story). So many people talk about 35mm film having massive potential
resolution, however, potential
are two different things.
This is also true when considering 4K cameras like the RED. Due to limitations in optics and imaging sensors, true visible
resolution tops out at around 3K or so. So, while 5120 x 2160 might be have potentially better PQ than 3840 x 2160 anamorphic, there isn't a camera in the world that will capture that kind of detail. Perhaps someday, but not right now. And 4K also seems to capture all visible detail from even 70mm film. Again, someday we might have camera systems that can truly capture these incredibly high resolutions, but for now, these high resolution figures are just numbers that have very little relationship to picture detail we can actually perceive.
No, we don't have any demo clips we can share of MFE due to NDAs with the studios we have licensed footage from.
If someone plays back an MFE disc on a standard display and player, the black bars are Java generated black bars that cover the additional resolution stored "behind" them. For that reason, they are totally opaque.
We have demo material with standard letterbox material on the left and MFE generated letterbox material on the right. No one yet has been able to discern a difference.
Hopefully I addressed the last question when I mentioned that MFE utilizes square pixels