For some media, this is absolutely the case. Particularly films that are slightly damaged, or seriously damaged, it's well worth holding onto the material in order to wait for better techniques to become available.

Case in point, this recent restoration of Morecambe and Wise by BBC R&D: http://www.bbc.co.uk/rd/blog/2017-12-morecambe-wise-video-fi...

They laser cut the film apart, then used X-ray microtomography to scan the film and some very clever software to unwarp layers from the 3D data produced by the X-ray microtomography process. Not exactly something we could've done 5 years ago.

Now that was an excellent read, quite astounding.

Please post to HN front page and re-post as updates occur!

BBC archivists also figured out how to recover colour information from black-and-white film.

https://www.youtube.com/watch?v=CjK-b4x9ZmQ

This is a great read about the restoration of a lost archive, but it is worth noting that in this case the source material was similar to what is being converted and destroyed now: analogue video tape recordings from a low-resolution source material. The source was recorded in PAL/NTSC/SECAM and the conversion process can be completely lossless and still end up with a copy that has every bit of information that was on the original.

In the case of the Morecambe and Wise recordings, this was video to film (with an additional step to try to lose information in the first conversion so that the scan lines were less visible) so the interesting technique here was simply in restoring a degraded source. Unlike scans of film negatives, there is no way a new process or technique is going to dig out information that the original low-fidelity recording system was simply unable to capture.

It's not uncommon to see vital documentary footage that was crudely digitized twenty or more years ago, to a resolution that was less than standard TV resolution at the time. It looked okay on TV then, but looks horrible now, and the originals are lost forever (since that was the point of digitization, to save rent money in the long run.)

Two examples from France: in 1998 AFP digitized the World Cup 98 pictures. In 800x600... Now they simply have no easy access to their '98 archive. From 1999 on, INA (French national archives) digitized videos in MPEG2 8mb/s 4:2:2 @MPML. That was supposedly "good enough" in comparison with the old media it came from (and was an admitted standard for archives worldwide back then). Actually, it was the only not unreasonably expensive solution, but in retrospect the quality quickly was deemed completely unusable and everything had to be digitized again in the past few years.

Yeah but film has grain and at one point your resolution becomes finer than the grains. At that time you know new technologies won't bring much more precision.

Now if the support is in poor conditions, perhaps the technologies mentioned in the other posts are worth considering. But if it is a tape or film stored in decent condition, it is unlikely that future techniques will add much to current ones.

That was the logic behind the choice of MPEG2 8mb/s @MPML: you're full (SD) resolution, full color, so you're better than the source (for instance, U-Matic is ~200 lines of ~250 points and a 4:2:0 color space). However, the switch to HD broke all expectations, because the low resolution of U-Matic captured in full SD and upscaled to HD is awful. Somehow that doesn't work out; you must capture even the very low resolution supports with a resolution as close as possible to the target work format (a so-called pivot format). Maybe upscaling the current HD JPEG2000 archives to 4K will fare better, and maybe they'll need to digitize everything from analogue once again...

It's not that bad if they're keeping the originals. Scanning in 4k HD in 1998 would not only have produced files too large to be processed and stored practically by computers of that time, but there was no scanning equipment capable of that. Also gives them a nice justification for their budget year after year.

In 2003 I did a consulting gig for the _Forum des Images de Paris_ who planned to convert their movie archives to digital. Back in early 2003, scanning the existing video and film stock in HD of any kind (720i/p or 1080i/p) was incredibly expensive and not generally available (most post houses simply hadn't any HD scanners or video capture cards yet, except for high-end film work); I unfortunately had to advise them to stick to SD...

But that would have destroyed the incentive - to save on rent for future decades.