Dec 18 2009
“Can We Enhance That?”
httpv://www.youtube.com/watch?v=Vxq9yj2pVWk
Because “enhancing” is to action/crime shows what “I can’t get a signal!” is to horror movies. The only one I could believe is Battlestar Galactica, where by the time we invent FTL drives, we might actually be able to enhance an image of nothing into something. But it still took TWELVE DAMN HOURS on that show.
It’s a flat image people! If there were more pixels to be “enhanced” you would see them already! Someone who knows science or computers explain this better for me.
More Unreal Posts
I have been talking about this for years. I have worked with computers for over 30 years and have done a lot with photography and video in the past.
I am not at all an expert in recent software programs and technology, but the basic concept is that photos and video are made up of pixels – tiny dots of different colors and shades of colors. Digital technology converts each dot to a digital code that can be stored and viewed.
Based on the quality of the camera taking the photo or video – how many dots per square inch – , there are only so many dots that make up a particular image or section of the image.
To say that you want to “enhance” the image, would mean where there are spaces between the dots, the tool would have to make a decision on what color/shade to make the “fill-in” dots. This could be done, but it would be questionable on whether it would be accurate.
It is especially silly in TV shows and movies where they take a surveilance tape from a 7-11 and “enhance” it to be able to read a tattoo on a guy leaving the store. Stores do not use high-quality cameras so the number of dots that make up the guy’s tattoo would be very, very small. Blowing it up to be able to read it would mean trying to fill in a huge amount of dots – virtually impossible to make readable.
Alrght, since we computer geeks have been called to arms:
You can’t increase pixel count. Once a shot is taken, it is what it is. Period. End of discussion, nothing else need be said.
Usually on these shows they “enhace” a picture to show a small detail. Uhhhh… that’s actually just zoom. But “Zoom That” just doesn’t have the same appeal. If something is pixelated and you zoom into it it will only become more distorted.
Sure, there are actual “enhancers” but these generaly are just to adjust the color/contrast in such a way that it becomes easier to perceive whatever is being analysed. Like say a number you can’t tell if it’s a 3 or 8.
NCIS will fail at this once every 3 episodes. They must have a patent on the “Enhace the image” trick.
Remember on Bad Boys 2? Dude took some shredded papers ran it through an “image enhancer” that decided where everything fits and then the program “knew” how to fill-in the spaces. Dude, do you know how advanced such a system would be? To be able to decide what artificial pixels to insert between the real pixels in order to enhace the image…. from basically confetti…. and incomplete confetti too since some pieces stayed behind.
Yeah, I’ve got no signal on that one.
@Bigdog
One in 3 episodes isn’t that bad of a ratio compared to some others.
Meanwhile… you pay to have an ultrasound of a pregnant lady, and we don’t know if the kid’s thingy is a realy dong or just his leg.
Let me stop now. lol
You guys totally enhanced my knowledge of this subject.
Hello,
In certain cases there are ways to “enhance” images, but you have to know something about the picture. For example say you took a satellite picture of the country side. But some of the data got corrupted so you only have a partial picture. However, you know something about the geography of the area or you have just taken a picture or you have a good guess. It’s the opposite of spam filtering. Spam filters delete mail because that mail looks like spam based on previous knowledge. (Spam filters are a lot better than simple networks because they have probabilities built into them). In reverse we generate a pixel of information based on the current pixels that we see coupled with the knowledge of what could go there. It can’t be done in the general case, but astronomers do this when taking pictures (see super resolution in wikipedia). Also see Bayesian networks.
If you know bandwidth of information then it is possible to do tricks but these tricks have to be done before taking the picture.
There is a lot of research in this with regards to MRI pictures as well. It will be never like that on tv, but information once lost can only be guessed, it can’t be guessed at.
I meant “can only be guessed at, it can be known for sure”
The movie Rising Sun is essentially this stretched out to two hours.
@everyone
Good work my geek reader army! I have learned much from your knowledge.
Others have summed it up quite well, you can’t see what’s not there. The photo groups small areas, pixels, and assigns them a colour that will look closest to the real image. This colour and it’s coordinates are then stored as a line of code. The higher the resolution, the smaller the pixels and therefore the greater amount of detail can be stored. That’s one of the reasons why old video games have such little detail, because the pixels were too large. By zooming into a picture, all that happens is the pixels now take up larger portions of the screen, still the same amount of pixels. There’s no feasible way to “enhance” it unless there is more data
People have already done a decent job of explaining how pixels work, however, though this isn’t even plausible if cameras used a better format we would be able to “enhance” them.
This is especially true for regular cameras, that can shoot pictures using a RAW-format which basically means that all the picture data is stored and with a simple light tweaking you will get a lot more “detail” out of shadowed areas. But enhancing like they do can’t be done!
As it annoys me every time they do it in shows, i just go with the idea its like google maps. ya know zoom in and it uses the pixels that wornt being used when zoomed out. with the wait for it to fix up image.
There are some things you can do to tweak the image, and especially if you have multiple extrapolate what it should be. But it’s really not quite the 15 second process they want it to be. It’s more playing with settings and tweaking things to get a best guess.
So, the answers they get are at best a guess at the truth. They’re just not going to end up with a crystal clear image of something, unless of course they take that shot separately.
I suppose it’s worth noting that there was an “enhance” seen in Avatar. Made me chuckle a bit.