|
First of all you need to understand how digital cameras record an image and how that differs from film:
CCD and CMOS
The CCD, or charge-coupled device, is the image-sensing device for most digital cameras. The camera lens focuses an image onto the CCD, which is a square-shaped array of tiny holes. Each hole is an electronic eye, which converts light into electric charge. The more intense the light is, the higher the charge. When the array is scanned, the charge is passed from one pixel to the next and read off the end, and each charge on each pixel converted into a number. The result is an array of numbers inside the camera’s microprocessor that corresponds to the intensity of light at each point on the array. The CMOS Sensor (Complementary Metal Oxide Semiconductor) is another image sensor that collects light like a CCD, but on a CMOS sensor each pixel is a charged plate with its own switch, and the charge is read off the chip one pixel at a time, in a different process to the CCD but one that yields roughly the same result: an array of numbers for light intensity at each point. In either case, the sensor is scanned repeatedly, at up to 60 times a second, building a history of images which are recorded to tape. On most cameras the CCD or CMOS array has about 400,000 pixels, in an array approximately 700 by 600. Because the array can reproduce a pattern of about 500 white and dark lines, it is said to have 500 lines of resolution. When I speak of “lines of resolution” I mean TV lines, or pixels, not the engineer’s meaning of the word which means line pairs. The CCD or CMOS chip is inherently a black-and-white device. There are two methods to getting color information. One of them uses three identical chips, each behind a colored filter of red, green, or blue. This type of camera is known as a three-chip camera. The other method uses only a single chip. With a single-chip camera, each hole in the array has its own, tiny, red, green, or blue filter. The one-chip method has worse color resolution than the three-chip method, but one-chip cameras are cheaper because there is no need to use a beam splitter or scan the extra two sensors. The chip has a limited range over which it can accurately measure the intensity of light. If the ray of light is too intense, the sensor saturates and measures the same value even though the actual value is higher. This is known as white clipping. Usually the clipping level is about a hundred times higher than the noise level, which is the lowest level the chip can sense without being drowned out by noise. It is noise which limits the low-light sensitivity. The noise is caused by random fluctuations in the electrical charge in the sensor. The ratio between the noise level and the clipping level is the dynamic range. Dynamic range and resolution are the most important characteristics of the chip sensor.
Color is created by using three layers of emulsion, each behind successive layers of color filtering which interleave the light-sensitive layers. After processing the filters are dissolved, but the red, green, and blue dyes remaining in the emulsions work together in separate layers to compose a full-color image. Film is also subject to a saturation effect, but the saturation level is not abrupt. Thus there is no “clipping," but a gradual reduction in sensitivity. It’s possible to overexpose film a hundred times higher than normal and still yield a somewhat pastel image of muted color. For the low end, film is also limited by noise, but in this case the noise is the grainy effect of the particles in the emulsion. The graininess of the film also limits the resolution, although very fine-grained film can register incredibly detailed images. It’s possible to photograph a newspaper and then read the text off the negative with a microscope (that’s how microfilm libraries work). It will be a long time before you can do that with a video camera. The resolution of 35mm film is several thousand lines.
However it’s rare for a filmed scene to use the full range of light. The reason is that most films are shown on television, and any scenes which use a five-stop range or more are not going to look very good on TV. Since a large part of revenue for feature films comes from home video or television, there is a great deal of pressure on cinematographers to shoot for acceptable television broadcast. And they do. I would say 99% of the scenes in a dramatic feature film could be recorded with a digital camera with acceptable results. The other 1% are situations where the white clipping of CCD’s or CMOS sensors produce a noticeable degradation of the image, but in later chapters I’ll discuss how to get around that. The resolution of 35mm negative film is about ten times that of a typical CCD or CMOS sensor, at least 5000 lines for film vs. 500 for SD video, or 1000 lines for HD chips. But the full 5000 lines of film do not really make it into the movie theater. Because the camera negative must always be printed to another strip of film to create a positive image, the viewer never sees the original film, but a copy, and often it’s a copy of a copy of a copy. Theatrical film prints are made from an internegative, which is printed from an interpositive, which is in turn made from the original camera negative. These printing steps are typically accomplished with a 1950’s vintage machine called a contact printer, which runs the film through at a blindingly high speed in order to reduce the time needed to make a print. The contact printer has no lenses or shutter, but simply brings the two pieces of film together in a gate, so the light must pass through one piece of film (the master) to expose the other (the print). The contact printer is not perfect, however. The two pieces of film are not tightly held together. The emulsions are not perfectly thin. There is vibration and looseness on the drive sprockets. Because of these factors, the full resolution of the negative is lost when the image goes through contact printing. The film projector in the movie theater is also responsible for a great deal of lost resolution. Unlike a movie camera, a film projector does not have registration pins which hold the film steady when the shutter is open. The projector uses a Geneva gear, an intermittent gear movement which is another piece of ancient technology. The steadiness of the Geneva gear is pretty bad, although it’s extremely simple and reliable. The consequence of poor registration is that the film shudders and shakes in the projector gate, smearing out the image on the theater screen. The projection lens and the projector focus also limit the sharpness of the image on the screen. The quality of theatrical presentation varies widely, but even the best theaters do not project with more than 2000 lines of resolution, and a more typical level is between 750 and 1500. My 1980 American Cinematographer’s Manual specifies a minimum of only 800 lines of resolution in their focus (depth of field) specifications.* So, considering the state of the art in theatrical distribution and projection, the resolution is not remarkably higher than digital video.
Television shows shot on film, however, which include many dramatic series and nearly all made-for-TV movies, are shot at 30 frames per second or 24 frames per second, on film. This is done mainly for two reasons: one, because film equipment used to be easier to use on location than video equipment (this was before the advent of compact digital cameras but the tradition has carried on since); and two, because shooting on film makes the finished project look like a movie. One of the main reasons for this “movie look” is the slower frame rate. Motion is broken up into larger steps, and the blurring in each frame is greater. The difference between 24 fps and 60 fps is very noticeable, even on simple close-ups. There was even a Monty Python sketch about this on their BBC television series, because BBC shows were shot on both video and 16mm film. The 16mm was used for exterior scenes and the video for interior scenes. It’s hard to say which look is “better,” but it’s sufficient to note that there’s a difference, and when you shoot NTSC for theatrical release, you have to take steps to ensure that the 60 fps to 24 fps conversion will go smoothly. If you use the proper settings on your camera, the 24 fps film version will look as if it were shot at 24 fps. Shooting PAL for theatrical release also involves some care. Also, there are many new HD capable cameras that can shoot at 24fps requiring no conversion or a simple pulldown removal. The information on these techniques is in Chapter Five.
Another advantage to digital sound is that with digital sound production software like Sound Forge, Sony Vegas and Pro Tools, you can create a theater-quality soundtrack with surround sound and stereo effects, using your home computer and a little bit of work. The results can be played using Dolby SR, any of the digital theater sound systems, or the Dolby Pro-Logic system for home video. All it takes is the software, a Pro-Logic (home theater) decoder, some good speakers and a good ear. Chapter Nine also covers this area.
For low-budget films where the primary goal is to make money from television and home video rather than theatrical release, and the only theatrical screenings are likely to be in film festivals, the slightly lower resolution of digital video, if anyone might notice, becomes a moot point. Assuming that digital video has no serious drawbacks in quality, what are its advantages over film? There are many!
If your budget is over a hundred thousand dollars (as it was with REDBOY) the difference of fifty grand may not seem significant. However, much of that money is spent up front, before you begin shooting, so the initial investment is higher. But if your budget is zero, like it is with many independent films, where everything has to be done as cheaply as possible, every dollar saved can mean the difference between completing your film and not completing it.
Publishing Update (14-Oct-2010): The latest edition is ready and we have them in stock!
|