Photography in Hollywood
6. April 2013, 13:28:32 Uhr:
Note: I've written the following years ago, but have never published it. So why not now?
Photography and Hollywood is usually a bad mix. And by that I don't mean the paparazzi problems, but how photography is presented in a typical Hollywood movie. A lot of really nice movies are made in Hollywood, but whenever it comes to science and technology, many movies fail. Whenever someone operates a computer in a movie, you involuntarily think “hmm, it never looks anything like this on my computer”. When movies touch topics like physics, they almost never get it right. Even if you allow for some stretch, often they are even ridiculously far from reality. You can't help but wonder what they meant when they said they had help from experts. These certainly didn't help to get the facts straight…
Photography is no different. Whenever you see someone handling a camera or working with film or video in a movie, half of the time you wonder if the actor ever held a camera before shooting the movie. Even if you're only a half-serious amateur, you know what you're shown is either plain wrong or far from reality, and you like to scream “hey, get a clue before shooting such a movie!”.
I bet you have seen all of the following at least once in a movie or TV show.
Blow it up!
This is maybe the most pervasive photography nonsense told over and over again in movies and on TV. Even a whole movie was made based on this imagination of some screenwriter. It usually goes like this:
A photo or piece of security camera video was found, and the investigators look at it on a screen. It looks unsuspecting, but then they spot some small detail, a face in the background or a faint reflection. Then they blow up ever smaller crops of the original image to huge sizes, and every intermediate step is razor sharp. In the end, they can clearly recognize the face or detail, and the story has moved a big step forward.
This is nonsense. The typical image source for these things is often a casual snapshot, security camera video or even a newspaper clip on microfilm. The technical quality of these images is low, to say the least. Once you enlarge it beyond the resolving power of the original medium, you won't see any more detail than before. You can do all you want, you can not get more information into the enlargement than what was already in the original. When the thing you're looking at is just an unsharp blob in the original, it will be just a larger unsharp blob in the enlargement. It won't become sharper, and it won't reveal more detail. You can vary brightness and contrast to a certain degree to make things more obvious. But this is nothing like what we see in the typical TV movie every week.
Let's do some math here: With a good film, a good lens, a tripod and proper technique you can expect a maximum resolution of about 120 lines per mm on film. That translates to about 6000 pixels per inch. When you use this as input and print it on paper at about 300 pixels per inch, you have done a 20× enlargement with a print size of 72 cm × 48 cm. You get exactly one pixel on the print for each pixel in the original. When you make even bigger enlargements, you will only get bigger pixels, but not more. You have exhausted the resolution of the original, and bigger enlargements won't reveal more details.
However, you will rarely achieve this level of resolution in the original. Use faster film, use a less than excellent lens, shoot hand-held, and the resolution will go way down. You will hit the limit much earlier, probably as early as with 6× to 8× enlargements. This is much, much less than what is typically shown in these scenes!
What annoys me especially about this plot device is:
- It's not new. It's repeated again and again since the 60s. Do screenwriters never get bored of it? I know viewers do. Whenever such a scene comes up in a movie, at least a dozen other movie-goers around me moan or go “oh no!”.
- You can see it from a mile before it's coming. Whenever the investigators gather around the technician's screen you think “please, not again!”, but inevitably the lead person will point to some blob on the screen and say “can you blow that up a bit!?”. You know exactly that after this scene they will have the case solved and the suspect arrested.
- It's unnecessary. There are more realistic ways to solve the puzzle, and the end result may not only be less predictable, but also more interesting. Why not try it this way?
2D becomes 3D
This is phase 2 after blowing up the image a hundred times. When the investigators still can't see what they're after, the technician will open his magic toolbox and will create a 3D animation from the image. They can then rotate the scene in 3D and do all kinds of neat things, like looking into a bag that was photographed from the side, or look underneath a cap that was covering a face.
Do they think we're dumb? You can't create a 3D model of a scene from a single 2D shot. It just doesn't carry the necessary information. You need several, if not dozens of shots from different directions to extract a half-decent 3D model of the scene. Even then, it takes a lot of processing power and manual post-processing to get something usable.
A photographer like a rock
We've seen it in dozens of movies and TV shows: A detective or investigator follows a suspect in a car, typically at night. He follows the suspect to a restaurant or café where he or she meets with other people. The detective quickly pulls to the side, sticks a camera through the car window, zooms in and fires a few shots when the suspects hands an object to the other person. Later in the movie, we see the prints from these shots. They're razor sharp, there's no camera shake, and the persons are perfectly stopped in motion.
The photographer in this scene must have a really steady hand! Streetlights at night aren't really that bright. I've done some test meterings. With ISO 100 sensitivity and at f/4 you get exposure times of about 4 seconds. Even if we assume that the detective is using really fast film, e. g. ISO 3200, the exposure time would drop to only 1/8 second. That's way from the 1/300 or 1/500 that you need to both eliminate camera shake and stop subject motion.
Sometimes producers like to make the detective look really sophisticated by letting him use a mirror lens. Mirror lenses are used by astronomers to take photos in the dark of night, so they must be ideal for the job, right? Of course, they're not. Most mirror lenses have a fixed aperture of f/8. That's two stops slower than a medium speed lens at f/4. This means they let in only 1/4 of the light that the medium speed lens would pass through. Using a mirror lens to take photos at low light conditions is exactly the wrong thing to do.
Needless to say that this plot device is often combined with blowing up the print to infinity to see what the suspect is handing over. Considering the circumstances in which the original photo was take, this is even less realistic than blowing up a photo that was taken under ideal conditions with proper technique.
High definition security cams
When you see images taken by security cameras investigated by the lead characters in a movie, they typically are of astonishing technical quality. They're razor sharp, high resolution (suitable for unlimited enlargements that will inevitably follow), noise-free, undistorted, and you can watch the recording like you can watch a movie.
Have the movie makers ever seen surveillance from a real security camera? It's low resolution (after all, it's only TV), the optical system of the camera is often low quality, it's noisy and distorted. Often a single video tape is used and re-used for years, so expect a lot of drop-outs. The system isn't recording 24 or 25 frames per second, but more like one per second, so that they can record an entire day on a single tape.
You can do something to enhance the image taken from such a video tape. For example, you can take the two interlaced frames of a video image and combine them to a single frame. This will reduce some of the noise present in the information and thus improve the signal to noise ratio. But no matter how much you do, you will not get even close to the quality of a rather bad photo from a still film camera. It will not be like what you see in these movies. You will not be able to enlarge it in a useful way.
Often the photo equipment seen in movies make sounds that you'll never hear from them in real life. I've seen Leicas producing the sound of an AF motor. I've seen cameras with manual film advance producing motor winder sounds – while the photographer was operating the winding lever! During the classic blow-up sequence you often hear the whirring sound of a small motor from the imaging equipment. When you blow up a picture on your computer, does it make a whirring sound? Sometimes it even sounds like a clockwork or gearing that is shifting optical elements around, like in some astronomical telescopes.
Making CGI look more real
This is not really about photography in movies. It just struck me when I first saw it. Sometimes in CGI sequences you see flare rings around bright light sources or ghosting opposite the center of the frame. The entire sequence is CGI generated, with no real lenses involved, so you know that no flare should be present in the sequence. Soon you learn that these effects were added intentionally to make the scene look more realistic, as if it was shot with a real camera.
How ironic! Photography people and movie equipment people are fighting hard to get rid of these unwanted imperfections of lens systems. We invent ever better lens coatings, use ever more sophisticated lens hoods, design lenses so that internal reflections are avoided or cut. Literally hundreds of thousands of dollars are spent to make flare a thing of the past. And now the CGI people add it artificially to produce a look that is more “realistic”. Think about it, without a camera recording the scene, these effects wouldn't be there in the first place. So “no flare” would be realistic.
These crime scene people are real pros
If you see a TV show like “CSI” or almost all of its predecessors, you'll notice one guy who is a real pro. He has a camera with a flash mounted on a bracket. He'll walk up to the evidence found on the scene, point the camera on the object, and shoot. The result is a perfectly lit and exposed, tack-sharp macro shot. Of course, it'll be used for limitless enlargements as described earlier.
If you did that, you wouldn't get a decent shot. At macro magnifications, DOF is really shallow. It's measured in millimeters, not meters. Move a tiny bit, and your subject will be out of focus. You will almost always use a tripod for this shot. It will allow you to focus precisely while keeping the exact same distance. It will also allow you to stop down to small apertures that will give you at least a few millimeters of DOF, and you'll not have to worry about camera shake.
At these distances, the parallax between the lens and the flash will be significant. With a regular flash bracket, the flash will illuminate the spot next to the object, but not the object itself. You'll need a bracket that allows some freedom of movement for the flash, and you'll have to adjust the flash as carefully as the camera. Maybe you'll even use the flash without a bracket, so you can light the object more from the side to emphasize the surface structure if necessary.
How to hold a camera… not
Whenever a movie contains a scene where a spokesman gives a press conference in front of a row of journalists, have a closer look at the photojournalists. You may be up for a laugh!
More often than not you'll see at least one guy playing a PJ who does the following:
- uses a camera that seems to be taken from a museum or attic, not from the toolbox of a current PJ.
- holds the camera with both hands left and right. He doesn't even try to support the lens.
- never even touches the lens, even though it's obviously manual focus and manual aperture.
- the lens is often inappropriate for the situation, for example a telephoto lens used at close distances in a crowd.
- almost never looks through the viewfinder. His look almost always goes over the camera directly to the speaker.
- presses the release at random intervals, sometimes not at all. Even though the camera is manual advance, he never operates the advance lever.
It's obvious what happened here: the film crew found an old prop camera, and they stuck it in the hands of a clueless extra who was supposed to play a PJ.
If you want to see how it's really like, watch the next news conference on TV. Maybe the movie producers should have done that, too.
Give it to me, baby!
Also seen quite often is a glamour photographer working with a model in a studio or outside. It goes like this:
- Photographer jumps around the model, never keeping one distance or even keeping still for the shot.
- Wildly turns the zoom ring or focus ring.
- Shoots either randomly, often faster than the studio flash can re-charge, or not at all (yes, I've seen this).
- Tries to pep-talk the model. Sentences like “yeah, give it to me, baby!” are often heard.
That's just silly, nobody works that way!
Specific movies and TV shows
Here are a few movies and shows that came up on my radar. If you haven't seen them yet, beware of the spoilers!
Filmed in 1966, this is the mother of all movies containing scenes where a photo is enlarged far beyond its resolution. In the decisive scene, the lead actor is blowing up shots from a film that others are very keen on getting. In the background of one of the shots, he discovers something suspicious. Inevitably, that shot is enlarged beyond belief to reveal the nature of what was going on in the background.
If you know the films available in the 60s, you know that enlargements of that kind are impossible, especially if you also consider the way the decisive shots were taken. The enlargements would have stopped to reveal more information long before the size of the final enlargements.
Enemy of the state
First, an ornithologist installs an automated video camera to observe waterfowl on a lake. The camera accidentially films a murder taking place on the opposite edge of the lake. When the guy discovers the murder, we're treated to yet another “blow-up” scene.
Then I saw the “magic 2D to 3D transformation” for the first time. The main character is filmed by a security camera in a lingerie shop. The bad guys then take this shot, transform it to 3D, rotate it and find that the ornithologist hid the video tape in the shopping bag of our hero. I had to brace myself so I didn't laugh when I saw this scene.
CSI, all the installments
These shows are a hotbed for all kinds of photography nonsense. While the main occupation of the lead characters seems to be looking cool and trying new sunglasses, you'll also see unlimited enlargements, free-hand macro shots, high-quality surveillance video or a combination of all almost every single week.