“Let’s Enhance”

16:29 Sun 07 Mar 2010
[, , , , , ]

A particularly inane trope:

In case the point isn’t crystal clear: you can’t do that. There are no image enhancement programs that let you know what the data missing at the point of capture is.

There appear to be plenty of people who think this kind of thing isn’t just possible but trivial (people who have “chosen to be smart at other things”, perhaps), at least judging by the amount of times it comes up on Clients From Hell.

I’m prepared to cut Star Trek and Blade Runner, and to a lesser extent Battlestar Galactica, some slack given that they’re science fiction, and we don’t actually know what the software capabilities or the capture format was. In Blade Runner, my interpretation of that scene is that what Deckard has his computer system analyze isn’t a simple photo, but rather some kind of hologram with a much larger amount of stored information than is apparent.

The other clips, however, deserve nothing but scorn and have no excuse. The fact that it’s doing the impossible is one thing, and then there’s also the fact that it’s a completely tired idea that exemplifies bad, lazy writing.

Put aside the first consideration for a moment and assume that this kind of technology is commonplace—you wouldn’t have characters reacting the way they do in these scenes. Either the person responsible for the analysis would simply do it themselves, or, nobody would be impressed at all by the fact that it’s possible. Instead, here you get a mixture, as if the technology exists but somehow the character is a total genius for thinking of using it.

Note that some forms of enhancement are possible; if you get a blurry shot of a license plate, it’s theoretically possible to use probabilistic analysis to get the letters or numbers, because you know in advance what the possible answers are. But that’s very different from what’s shown in those clips.

Also note: this post is not meant to imply that any other aspects of the shows referenced above are necessarily more realistic in any way.

4 Responses to ““Let’s Enhance””

  1. Steve Casey Says:

    In principle you’re right – you can’t get more info out than was put in. However, almost every example does talk about enhancing a still from video footage. Now that’s not so far fetched. A still only represents 1/25th (ish) of the information per second on-screen. You, in theory, could take the other 24 frames and work on mining information from them to contribute to the still image.

    It’s probably not in the scope of today’s technology, though look what amateurs can do by combining multiple images/video to create celestial photography from smaller scopes.

    On the other hand, the Blade Runner one was outlandish, though sci-fi gets a free pass on IMHO.

  2. Kevin Teljeur Says:

    That ‘Clients From hell’ site is amusing. I’ve heard so much of it myself. And sometimes, just sometimes, I’ve delivered (the trick is to humour the ask, gently lower the expectations to below what you can deliver, and then surpass the new, lowered expectation. Professional design is about 10% actual design, and 90% dealing with the client.).

    I thought about this image enhancement thing too, having been a graphic designer and Photoshop specialist in the past, and of course every time I see this happen in a series or movie I shout at the TV, because I know that you can’t get out more than you put in. You can, by and large, only cleverly remove stuff.

    But the Bladerunner scene you’re referring to is different, because the technology is extrapolating unseen images from reflected light, etc – in other words, it’s suggesting that there are ways to build what’s in the image into something more. I actually like the idea very much, and I’m sure it will be possibly in the future; it’s at least creative and not very overblown. Similarly, Star Trek and Battlestar Galactica… Well, in the first, anything goes, and the latter, I seem to remember that it was still more believable than an average episode of CSI, especially considering that they’re battling clones and robots and jumping between star systems. CSI really takes the biscuit but we all know that. I’ve always wondered what people in forensics and medical professionals think of that.

  3. Graham Says:

    You’re absolutely right, Tadhg.

    Steve, what you’re saying about calling in help from neighbouring frames is theoretically quite interesting – indeed it’s very similar to how mpeg compression works – and Kev it’s true that sometimes the characters are using technology from the future. But frequently what we’re watching is meant to be contemporary and that’s the problem.

    The fuckers seize upon one frame and immediately leap into this fantasy world in which benign rows of differently coloured dots have magical properties. No, they are not just rows! Behind each row is another row and then another – and if you pivot left or right there are countless more rows. That’s really the most charitable explanation for the ridiculous visual revelations that occur…

    It’s not just that it’s impossible in real life. It’s that everyone knows it’s impossible by now. Maybe twenty years ago a scene like this, served up to us once, wouldn’t have been a problem for our mass suspension of disbelief. But nowadays, so many of us deal directly in the currency of megapixels and image enhancement. A shocking number of us know it’s bullshit. Yet they are still incorporating these moments into cheap drama.

    What really galls me is that it’s not even being done out of necessity or practicality – the familiar moment is being written into telescripts and screenplays because writers think it makes for a dramatic moment!

    There a lot of things you can say about it but I think Tadhg describes it best as “lazy writing”.

  4. Steve Casey Says:

    In case anyone’s still watching, no idea if there is any geocoding on their website, but a pop-view of the above subject. Some quite interesting things though only a very superficial review of them…


Leave a Reply