Is anyone anywhere attempting to digitally expand fullscreen footage to widescreen format?

It strikes me that the technology probably exists now to be able to rework fullscreen format footage into a widescreen view - not by cropping to a widescreen rectangle smaller than the existing frame, but by techniques such as:

  • Digitally adding in plausible scenery (either just made up scenery for static shots, or captured digital renderings of stuff that moved off the edge, in panning shots and similar)

  • Re-compositing the action so that it doesn’t just occupy the middle of the newly-widened scene.

Is anyone attempting anything like that anywhere? (I mean, for footage that has never existed in a widescreen format - obviously for films that have been P&S’ed, the proper solution would be to go back to an original print)

I don’t think the effort involved to do it that way is worth the return.

However, a lot of 4:3 footage for TV was on film, so can be rescanned into digital, reframed as 16:9 and still maintain the resolution.

I’m thinking mainly of classic TV comedy stuff - I don’t know if it was shot on film, but if it was, the stuff in the cropped edges probably isn’t interesting, or might even contain boom mikes, etc - it seems to me that with the prevalence of widescreen TVs now, there would start to be a demand for much-loved originally-fullscreen content, but without the black borders.

Is this what they’ve done with some mid to late 90s shows? I’ve been seeing Seinfeld, Everyone Loves Raymond, and a few others that are now in 16:9 format. I just wondered if they were using the older stuff or just forcing the 4:3 out.

I’ve noticed that too, and I’m not sure what’s going on, because I thought those shows were shot on video tape.

They just went and cropped it.

Or stretched it.

Face it - people who are bothered by black borders aren’t bothered by a stretched image for some reason. The technology already exists to do that - so why go to the trouble to create fake borders or go back to the original film?

Both things bother me.

I think it’s difficult enough to convincingly pull off what you describe on a single image in Photoshop. To automate the procedure so that you could handle hundreds of thousands of video frames would therefore be quite a challenge.

Dunno - would it be all that different from the kind of digital compositing that happens in movies right now? (i.e. adding digital characters or scenery?)

All widescreen TVs have a Fill or Stretch or Automatic mode that does this on the fly. Most of them are good enough that, aside from some noticeable noise because the image has changed sizes, actually looks pretty good. Very little gets out of frame. Except for those “big block of text” commercials that just don’t look right. But who cares about those?

There was a lot of old TV footage (news, sports and coumentaries – not studio shows) that was shot on 16 mm film. In fact, almost all kinescoped programs were shot on 16mm, rather than 35 mm.

When I worked in TV at the end of the newsfilm era, ALL our cameras were 16mm. Hell, I don’t think we even had the capability for 35mm anywhere in the station.

One reason, of course, was cost, but the second reason was that the aspect ratio of 16mm is 4:3 – the same ratio as standard TV.

A lot of the survivng footage may not be available in 16:9.

But not all of them will do it for all broadcasts. Mine will not do it for any x.1 broadcasts. I have to have it in whatever format they send out.

I’ve been wanting to ask about the 10-15 year old shows I’ve been seeing, it’s hard to tell, but a lot of stations seem to be broadcasting them that way. Sometimes they look like they were shot in HD, but that could just be because the TV looks a lot nicer now then it did 10 years or so ago. I never watched a lot of those shows so I don’t know what they originally looked like.

A lot of the older shows were shot on 35mm film which is better than HDTV today. “Hogan’s Heroes” is a good example of an old show that was shot on 35mm and put on High Def.
It was done as a “test case” so to speak to see if a declined old rerun could find new life it was made into high def.

It went over but not enough to make it profitable to do it to older shows.

“I Love Lucy,” was also shot on 35mm, a lot of shows where 35mm. Video tape became common in the late 70s and early 80s.

“Scrubs” was shot on 16mm film for awhile then it swiched over to 35mm, I’m not sure when, but it shows you can transfer 16mm to high def if done carefully.

One of the issues with the new high def shows, especially comedies is they may be in high def, but the extra space is wasted. Because so many people didn’t have high def, when they started shooting it or even now many people “crop” the high def to make the 16:9 fit 4:3, what the directors did was make sure all the action happened only in the area that would fit into 4:3, so the rest of the space on TV shows is just background and serves no purpose.

Even the station logs (called “bugs” that appear in the lower right hand corner have shifted over a bit so if you crop the screen you’ll still see the “bug”

No- it would be just as expensive, too. I mean really, does Paramount really want to spend hundreds of thousands of dollars digitally adding DJ’s artwork taped onto the fridge in every frame where Roseanne & Darlene are sitting at the kitchen table?

Do you know how much effort that takes? The amount of labor necessary to realistically render a background is way more than it is to create sets; even more so for foreground objects or characters. The only reason it is done is to allow actions that couldn’t be done with live action effects.

It is also the case that the director may have blocked the staging or location so that activities and equipment for filming are just outside the sightline. Improper matting in film often shows microphone booms, shadows, prop handlers, et cetera. I would especially expect this to be true in “three camera” setups that are often used in live studio filming, where the cameras may be arranged around two actors who appear to be facing directly at each other such that the camera sightlines are just over the shoulders and just out of frame of each other. (If you watch the director commentary on Heat about the “A cup of coffee” scene with DeNiro and Pacino, Mann explains that he used this setup to get the take “live”, i.e. the actors interacting with one another. Because of this, some people claim that the two were never on-set together, although it is clear in the wideframe view that they are actually facing each other.)


I’ve been thinking about this for a while.

It is definitely possible with situation comedies. A show like Hogans Heroes or Seinfeld exists on a handful of sets, and virtually no exteriors. Every part of those sets exist in the library of videotape or film, and gets added to the database for Image Based Rendering (the IKEA scene in Fight Club was an early example of IBR). The sides of a scene become available to be patched on. Scenes with other people entering from one side get recomposed with more virtual set to the other side.

It’s a technology that’s been available as a research tool since the mid-90s, and should be a standard tool from a company like Quantel in a few years. It will probably be semi-automated, with an operator tweeking the obvious CG foul-ups.

Why on earth would you want to do that? That sounds like something that people who want their black and white movies colorized would do.

Hogans Heroes and Seinfeld were my examples, not Citizen Kane.

I prefer OAR. I hate stretched images and I hate cropped images. I’d rather deal with black bars on the sides and see what was intended, rather than actually lose info or distort it. The only time a 4:3 image is bad on a 16:9 tv is when the black bars burn into the screen. I’ve had that happen before.