I recently watched an episode of Law and Order. They were examining a video tape where two guys were carrying a plasma TV set. In a point some of the cops saw a sticker in the crystal of a car, but the detail was blurred. The operator selected a portion of the frame, then zoomed in it and with some strange computer-aided maneuver, the blurred portion turned to a perfectly clear image.
They couldn’t zoom more, but I can’t believe this can be done. Computers or any video technology just can’t invent information they don’t have.
Not only in TV shows I’ve seen this, in many movies, too.
You can do this if and only if the video is of a very high quality - in which case you’re simply zooming on what was already recorded. There is no software method of “sharpening up images” like that. And frankly, business do not invest in super-high-def cameras and reording systems.
I wouldn’t doubt that there are some sophisticated tools out there that have algorithms that make an ‘educated guess’ on what color pixels to insert to “create pixels where there are none”.
But I doubt they’d be too successful, and a crystal-clear image, forget it.
They can get more detail than what is evident in stills if they are dealing with a moving object and have several frames for comparison, but if you are working with a still picture, you can’t get data that isn’t there.
To continue rexnervous’s comment: I thought the JPL and NASA had software for sharpening images sent satellites and spacecraft. From my understanding, they have to work with a comparatively few number of bits, so use software to “extrapolate” pixel properties based on the properties of the pixels surrounding them. For instance, if a pixel is between a “white” pixel and and “black” pixel, then it must be “grey.”
Or something like that.
If so, I don’t see why something comparable wouldn’t be generally available. I agree, though, that how technology is portrayed on TV and movies is generally laughable, the product of some director’s imagination.
I don’t know about video, but if a digital image is high res enough you can get some surprising bits of info out of it. Frex, I used to write captions of people at meetings for a publication I edited. I wouldn’t have the nanmes of everyone in the photo every time. But if their name tag was visible I could sometimes zoom in on it and read their name quite easily, even though it was only visible as a squiggle when displayed on screen.
My exprience with doing screen captures from videotape on my computer is that video images gnerally lack resolution and depth. But that could be a product of my video capture card, which is an older one, rather thna a reflection fo the actual data present in videotapes.
. . . unless that pixel should be red, or chartreuse. And educated guess would still not be good enough; AFAIC that’s how the photoshop sharpening filters work, but they don’t do much.
Bottom line: if the pixels are there, you can’t create them out of nothing. So, the OP was correct in calling bullshit.
Yes, it’s nonsense, but that sure doesn’t stop just about every cop movie and TV show from using it as an integral plot point. I had some respect for CSI until they used the magic zoom to see a killer’s face in a tiny reflection off a corpse’s eyeball.
The same happened with the swapping heads of Sean Connery and Wesley Snipes in Rising Sun, as some doper commented recently. As if the program knew what was behind them.
IIRC, the time when that movie was made (1993), image processing software was not very popular yet. So people was amazed by what a couple years later was going to be possible in her own desktops.
It was used as a joke in the much-under-rated Super Troopers. Jay Chandrasekhar played a Vermont highway patrolman and was using a computer to pull an image from a dashboard video camera:
Jay: [fake english accent] Enhance. [taps keyboard five or six times]. Enhance. [taps keyboard]. Enhance.
Brian Cox: Just print the goddamned thing!
My favourite example is ‘Enemy of the State’, when they don’t just enhance the image, but also rotate the camera through 180 degrees. They say something about the computer predicting what is on the other side. Why bother filming it in the first place? Just have the computer predict what happened.
Anytime a movie professes knowledge about computers, religion, law, or science, you should assume that they don’t know what they’re talking about. You will rarely be proven wrong
I’m not sure which drove me more berserk: TV shows that did this (X-Files, I’m looking your way), or my ordinarily super-skeptical friend who would swallow it hook, line and sinker. “Well, with a good computer program, they could probably do that.” No they frickin’ couldn’t! I’ve worked enough with processing and editing photos, both film and digital, to know that there’s no program that can take a mall security video (bare-bones quality to begin with, plus tapes that have been recorded over X dozen times) and enlarge it enough to read the gray-on-gray lapel of a man sitting 50 feet from the camera.
The magic zoom wouldn’t bother me nearly so much if it weren’t inevitably a major plot point.
As a few posters have said, it all depends on the resolution of the source image. There are techniques that can sharpen or otherwise enhance an image, but the resolution is the key. For example, the CSI example about seeing the killer in an image of the corpses eyeball is possible if the resolution is good enough. I make this statement confidently since I have zoomed in on photos taken with my Minolta S414 digital camera and seen easily identifiable images of myself in the eyeball of my six year old son.
I know I’ve done little to definitively answer the OP, but I couldn’t resist such an ideal opportunity to share my eyeball photo story.
