The first one is easy enough. Given that 480p, 720p and 1080p all look best in their native resolution without any scaling, doesn't it make more sense that the respective media is mapped natively to the exact pixels? Specifically, shouldn't 480p content be windowboxed? Granted it would make the image smaller, but the resolution would be sharper, and presumably the TV setting could be adjusted to show a scaled image to fill the screen, in much the same way a 4:3 image can be stretched now. But on my 32" 720p screen, it would make the SD image about the equivalent of a 19" TV which was typical of the image size most people watched these shows on originally.
Part 2 of this observation is that most DVD transfers are of the full TV transmitted frame, and the entire overscan area of the original broadcast image. In some cases, as much as 15% of the image could have been obscured by a television, especially in the 50s and 60s. And at least 10% of the image was obscured by most shadow-mask TVs into the 90s. Given this reality, and the fact that directors and DPs used a ground glass with action-safe markings to frame a scene, what exactly constitutes the OAR for most programs produced for TV? Granted the aspect ratio is 4:3, but how much of that should be transferred, especially for Blu-Ray? I mean people don't buy BD to use with 4:3 CRTs, and albeit LCDs have some overscan, but nowhere near the amount as a CRT TV for which the material was originally composed. Moreover, most of these TVs the overscan can be turned off, and I would venture that most BD collectors not only know how to turn it off, but make sure they buy a set with 1:1 pixel mapping ability in the first place.
I first noticed this phenomenon watching a DVD transfer of Irwin Allen's "Time Tunnel". In one scene, a bunch of soldiers run by the tunnel, the back of which was not masked, and they bunched up against the back of the set, accordion style, and then stood there waiting for "cut". Obviously, they would have been visible in the TV aperture, but the director knew they would be concealed in the overscan area. In Star Trek TNG, CBS used an example of a shot that had C-stands, lights, etc. in the full aperture so as not to be able to convert to 16:9. Interestingly, when viewing the DVD TV aperture, used on the BD, the C-stands and lighting flags can be clearly seen on the edge of what would have been hidden by the overscan area.
Given this reality, shouldn't studios be trying to preserve the intended framing in these transfers? Clearly this overscan material was never part of the original framing, it was meant as a protection in case more than the action-safe image was seen, rather than black bars (see even in the days of 4:3 TV this was a problem). Most of the old TV series I see broadcast in syndication today crop the frame to the action-safe area. This has the effect of bringing the subjects closer to the viewer, even if framing is tighter -- this is what was intended, something TV has always done to compensate for the relatively small viewing space compared to motion pictures, and what was intended.
Below you'll see examples of the full TV transmitted frame as offered by the BD transfer, the adjusted action safe frame as composed by the director and DP, and the way it would have appeared on a typical TV from the 80s (hard to believe we ever watched TV that way). In addition to the distance removed, notice all the extra space surrounding the focal point (which would be the speaking actors in the center). In particular, there is so much bright overhead lighting as to draw the eyes upward. That surely was never meant to be part of the composition, nor the space underneath the front chairs. I used TNG as I had been playing with this with respect to that thread discussion. But obviously this applies to all BD transfers. (FYI, the 80s TV recreation was based on aligning a 480p test pattern with that appearing on the TV itself, then matching the BD image size, so this is accurate of well over 10% loss of image on a typical 80's set)
At a minimum, shouldn't there be a way to adjust the framing from within the BD software in order to get this kind of precision cropping? I realize there would be a minimum loss of resolution at 1080p, but certainly less than up-resing a 480p or 720p frame to fill a 1080p screen. Or how about a simple window boxing matte you could turn on and off? Yes you could do this on your TV, but they zoom features really aren't very sophisticated, designed for little more than making the sides of a picture hit the sides of your set, by whatever brute force necessary.
Many people are still freaked out about the letterboxing (or pillar-boxing) which happens when the OAR doesn't match the TV that they have now. Having more scaling options directly available to the user would be nice, but I suspect few people would care enough to learn how to use them meaningfully. Current readership excluded, of course!
Marantz SR7009 avr + MM9000 amp --> Atmos 7.1.4
Fronts=NHT 2.9+AC2, FH+TM=DefTech PM1000, LCR+TM amped
I dunno, the action is right of center and those are all little bits at the left edge. It's not something I'd see without it being pointed out.
Do you have a timecode for when this scene happens in the episode? I'm curious if the framing is identical to their example shot.
Precise under- and overscan (zooming) options are available in dedicated video processors and I'm using them. With the VP it's possible to get window-boxing all around: my VP can do something like 20% each side. Not quite down to a 640x480 frame but that would be too small for most people. The danger of course is image retention/burn after prolonged viewing.
As I see the whole frame, I've found in 4:3 SD broadcast TV material the size of the frame varies from program to program (with varying amount of black lines at the edges), there really is no single overscan setting that suits everything ideally. How would the consumer know how much to apply for each program to match exactly the director's intention? And to adjust the frame from program to program isn't something many would be inclined to do.
Audiosceptics accept audio trials using 25 people. A recent Oxford study with over 353,000 patient records from 639 separate clinical trials shows for every 1,000 people taking diclofenac or ibuprofen there would be 3 additional heart attacks, 4 more cases of heart failure and 1 death every year.
I don't think that is a given. Granted that scaling cannot create detail that does not exist in the source, good interpolation can produce a convincing illusion. I would have to be convinced that it damages the image.
Secondly: standard definition video (as on DVD) uses non-square sampling: the storage and display dimensions are not the same. It has to be scaled in one direction or the other.
I've found in 4:3 SD broadcast TV material the size of the frame varies from program to program (with varying amount of black lines at the edges), there really is no single overscan setting that suits everything ideally. How would the consumer know how much to apply for each program to match exactly the director's intention? And to adjust the frame from program to program isn't something many would be inclined to do.
In any event, my preference would be for the standard settings present on a commercial set, just as they are. Basically 5 settings I have commonly seen: 4x3, 16:9, uniform zoom (4:3to16:9), vertical stretch zoom (though I have never used nor wanted to use it and don't know why it's there), and just scan -- though there might be better choices. And these need to be physical buttons on a remote, not a step-through affair, for quick access (I suspect they are they way they are since manufacturers figure consumers need to see the effect on the picture and won't be able to figure it out by name alone). Then, from the TV menu, a set of advanced settings, that really only people like us would use, giving several more options. It wouldn't be the first time a consumer product had dozens of settings the average buyer didn't need nor knew how to use. Nevertheless, a set of fine tuning options like: enlarge picture 10%, 15%, 20%, with options to maintain aspect ratio or allow the sides alone to expand (but not stretch). I am constantly adjusting my TV screen based on the program. Many 16:9 broadcasts are still window boxed in a 4:3 presentation, and I enlarge it to fill the screen. Some are stretched and I return it to 4:3. Ocassionally there will be digital garbage along a border so I have to take it off just scan. Another set of options would be to maintain 1:1 pixel ratios, so that 480 and 720 picture would be window boxed to the appropriate pixel mapping, when applicable. Not that there are many 720p sets being made at this point, but I would kill for this function and live with the narrow window boxing around the frame of my picture rather than slight up-scaling to fit the standard 720p 1366x768 display. The point being, this is a software fix which shouldn't necessarily cost the manufacturer a lot to add, even though most consumers may not necessarily use it.
Remember, the original FX for 80's and 90's era shows were done in an SD video environment, meaning the output source was often 1" video tape. SD tape formats from that era tended not to cover the whole dimensions of the TV signal area, often leaving black lines on the left or right. This extra black was almsot always covered up by TV overscan, so it wasn't an issue. When you see the material without any corrective zooming on a 16x9 TV, you see that extra black.
Plus, for TNG, they often used what amounts to a control room style switcher to key in elements. Switchers often have what is called an effects preview on them, where a little sliver of the image in the preset position (the next shot you want to switch to) appears at the far left of the image in the overscan area. Control rooms use underscan mode on their monitors, so the person operating the switcher would be able to see that little sliver of video. It allows him or her to verify the next shot is really there and ready to go before transitioning to it. In the case of editing Star Trek, they might not have a video source in preset, meaning black would appear in the effects preview instead of another shot.
So, in theory, you'd end up with the following:
1) Shots with no effects and only standard color correction would appear almost full frame, cropped as necessary to get the proper composition and eliminate unwanted elements.
2) Shots with minor effects have black from both the switcher and the tape they get sourced to.
3) Re-used shots from video sources that have additional effects added to them (or background elements they use to layer additional effects onto) might have the black on the edge from #2, plus additional black from being put through the effects cycle again.
Every generation results in "image creep" where that tiny black edge gets bigger as the shot gets reused and changed. Plus, when they reframe a shot in post, sometimes you catch a bit of the barn doors around the camera lens used to block lights and prevent flares - so you get a black line or sometimes just a black corner from that.
Regarding the original post, I think any benefit of not upconverting the video to a TV's native resolution would be negated by having to view a postage stamp sized image on a screen - something that would be aggrevated by larger resolution screens. So, 480 lines on a 720p screen = small. 480 lines on a 1080p screen = annoyingly tiny.
As far as whether we should see the full image that includes the overscan area or just what audiences originally saw, I'll take the full overscan image since it makes the media more versatile by letting me view it on an older set (via downconversion) without excessive cropping from overscan.
Finally, as far as seeing occasional items you shouldn't see, that happens all the time. From boom mic shadows to parts of dollies and camera mounts, there are plenty of things we see that we shouldn't - and that doesn't even account for reflections. I saw a lot of those things on older shows back when I watched them on a tube TV with plenty of overscan. Sometimes, the folks making these shows just fail to see stuff at the time they are shooting and there's nothing they can do about it later. Sometimes they miss stuff in post, so they fail to correct it, even when they can.