So i’m watching CSI:- Miami over the weekend.It’s the one with the mad sniper on the rooftops. At one point they recover an ATM picture of a man, they figure out he’s looking to the side at someone next to him. So they go in for a close up of his glasses.There’s something reflected there. Horatio asks the tech guy to enhance it and Ta-da! there’s a clear picture of a badge which is later used to identify a guy.
This isn't unprecendented either, i've seen similar acts of photographic legerdemain in several previous CSI episodes. Now my gut reaction is, this is nonsence. You couldn't possibly recover so much detail.On the other hand it's amazing what we can do these days.
So anyone know the straight dope? Can photographic experts really enhance a grainy image that much or is my instinctive cynicism correct.
It depends on the resolution of the ATM picture. I once borrowed a friend’s digital camera (a CyberShot, if you must know) and took a picture of the city around me at the highest resolution possible. When I zoom in using Photoshop, the amount of details seem is quite amazing.
So if the resolution is damn high, it is well possible.
If ATM pictures are akin to security camera video, then no such luck. The pixelisation will be too great to make out any detail - the face will be just one big rectangle.
I think it’s part of a subtle message that is being given out which is supposed to lead you to believe that cameras and police can do this. Obviously it doesn’t hurt plot development either.
If it was possible, we would have crystal clear pictures of the bombings in Baghdad and Crimewatch type shows would have broadcast quality videos of people raiding petrol stations or 7-11s
Also most people who operate Security cameras recycle the tapes that are used to record the pictures way past their lifespan so you have even less quality pictures stored than the camera is capable of.
I don’t think ATMs routinely take pictures anywhere near that level of detail. Its much closer to the level needed to look at the face of punk who just used the stolen card.
As i said it’s not just ATM’s. Another example springs to mind.From “Let the seller beware”. The one where they find the husband dead in the pool and his second wife shot dead.
Here they suspect the involvement of the first wife. Eventually (after the usual plot twists) they find a suspect for the actual killings. They prove his connection with the first wife by means of a picture from the 1st wife’s house where in the background there’s a framed photo on a table.
They then blow up the image of that old photo within the current photo to the point where him and the wife together are clearly visible. To which my reaction was. No way!.
I think Hollywood came up with that imaginary word “enhance” where some computer geek clicks a few buttons and suddenly the fuzzy image comes up crystal clear.
It was used in tons of movies and tv shows. I think I first saw it in Bladerunner.
I’ve done alot of digital artwork and seen some amazing stuff. But what you’re seeing on CSI I believe is hoky poky. If the resolution is poor and the information is not there, then it cannot be enhanced to the point of recognition that they are claiming.
I think this device hit its limits on She Spies, which admittedly is a total spoof. A kidnapped girl was shown sticking out her tongue at the camera. She had a tongue stud. They zoomed in on the ball of the stud and enhanced the reflection off of it to reveal the name of the company at which she was being held.
I see this all the time–watch X-Files for a while–and it’s bullshit.If the pixels aren’t there, they just aren’t there; no amount of “enhancement” is going to create them.
Yeah it’s crap. Keep in mind too that when somebody is “enhancing” a photo by blowing it up and trying to fill in details, the details have to be added by the person doing the work. As lissener said, if the information isn’t there in the first place it becomes conjecture. One person might see a full head of hair where another sees a bald person in a deep shadow.
Every time I see that happen, when somebody blows up a picture to something like 10,000 times original size and theres a bright crisp detail, I just want to beat the writer over the head with my hasselblad while saying “resolution,damn it, resolution!”
Yeah it’s crap. Keep in mind too that when somebody is “enhancing” a photo by blowing it up and trying to fill in details, the details have to be added by the person doing the work. As lissener said, if the information isn’t there in the first place it becomes conjecture. One person might see a full head of hair where another sees a bald person in a deep shadow.
Every time I see that happen, when somebody blows up a picture to something like 10,000 times original size and theres a bright crisp detail, I just want to beat the writer over the head with my hasselblad while saying “resolution,damn it, resolution!”
Recently I saw one where they were trying to figure out the identity of a person using stills from a security camera. The person was wearing a wig and large sunglasses, and had the collar of his coat pulled up. You could basically see the nose and some of the mouth.
The CSI guy yammers some technobabble about facial proportion mapping or somesuch, and then copy and pastes the picture of their lead suspect onto the image, and everyone gasps and says “that’s him!”. No duh, they could’ve made the person in the picture be Harry Potter if they wanted to.
I felt like I was watching a show about police set in George Orwell’s “1984”. Even if the person isn’t guilty, they can make him guilty if they want to.
It was done in “Enemy of the State” as well. In that movie, they had a security photo showing (for example – I don’t remember the exact details) what Will Smith’s left side. But he was holding something in his right hand.
The investigator instructs the computer person to rotate 180 degrees to show what was on the right side.
Um. People? Videos and photopgraphs are two-dimensional images. You can’t just flip a photo over to see what’s on the back of the t-shirt… Dumbest thing I ever saw.
It is possible to do amazing things with photographic material, and modern technology makes all sorts of manipulation and enhancement possible. But on CSI and other shows, they take it way too far. Sometimes, they take a legitimate principle (enlarge, enhance, filter, clarify) but stretch it beyond all credibility. At other times, they just make up crap because they can and it’s a story, damn it. As many posters have pointed out, the info you want is either in the pixels or it isn’t. If it isn’t, no amount of enlargement, enhancement or scientific woofle dust is going to help.
Not that I care on whit. I just want to see Marg Helgenburger, beautifully backlit with her lip gloss flaring ever so slightly, strapping on those white latex gloves and getting to grips with the evidence. It’s what TV was invented for.
Dogzilla’s ‘Enemy of the State’ example is a classic. A neat enough little thriller, I thought, but the 3D manipulation of a flat image was silly.
Sadly, I think this is a case of art imitating life. Remember the Unibomber? Remember how all of the police sketches just showed a vague-looking guy in a hooded sweatshirt? And do you further remember how, when Kazinski was caught, the news kept on showing clips of the sweatshirt sketch morphing into the completely different-looking Kazinski?
That’s the first scene I thought of upon opening this thread also. I saw it and kept thinking to myself “now is this really feasible?” I tried to convince myself that maybe the gubbermint had some super-duper computer running a top-secret algorithm that could figure out from minute alterations in the shape of the bag over time what was actually in the bag.
Well, unless the camera is storing the picture using a fractal algorithm, you just get big squares if you zoom in.
I actually thought enemy of the State did a better job than most, they bothered to make it clear that the computer can INFER what is on the other side but not definative. And they didn’t rotate it 180 degrees, it was more like 70,
That’s really weird, because later they make a distinct plot point that they can’t tell it’s Gene Hackman since he’s careful not to look towards the sky, where the satellite cameras are.
The picture super-enhancement schtick goes at least as far back as MacGyver. I remember an episode where Mac had some guy do the “enhance the picture beyond the pixels” BS. They called it “bitmapping” and gave some technobabble involving the X, Y, and Z directions in the picture.
True, but the sketch turned out to be useless. They only caught the guy because his brother read the manifesto and told the cops that it looked like Ted’s work.