There has been much written in this publication (and others) about the wonders of the new tiny HDTV camcorders, both HDV and P2. Unfortunately, there has been practically nothing written about the pitch of their sensors.
The what of their what?
In the camera part of a camcorder, there are one or more imaging chips. Each is divided into many thousands of individual photosensors, and the distance from the center of one to the center of the next is the sensor pitch.
Sensor pitch is almost never mentioned, but that tiny value can affect the amount of lighting required, the depth of a set and how detailed the pictures can be. Of those, sensitivity is easiest to explain.
If you had a photocell that, under certain lighting conditions, was putting out one volt, and you covered half of it with opaque material, its output would drop to half a volt. Cover half of the rest, and you’d get only a quarter of a volt.
In a 2/3-inch format, 16:9 aspect ratio camera imaging chip, the photosensitive surface area is 5.4 mm high by 9.6 mm wide. For American SDTV, there are not quite 500 rows of sensors on the chip. That means the vertical pitch of the sensors is about 11µm. Although the number of sensors per row varies greatly, assume it’s the same.
A 1/3-inch format imaging chip is about half as high and half as wide. That makes its sensors about a quarter as sensitive. What could be seen at f/16 on a 2/3-inch format camera will require either four times more light or an aperture of f/8 on a 1/3-inch format camera.
Then there’s HDTV. In the 1080-line format, there are more than twice as many rows as in SDTV. Going from 960 x 500 to 1920 x 1080 chops the sensitivity to less than a quarter, too. That means needing a total of 16 times more light or going from f/16 to f/4. Sony halves the number of sensors horizontally in its HDTV camcorders to double their sensitivity (at the expense of some detail resolution).
The depth-of-set issue is based on depth of field, a measure of how much depth is in focus at the same time. All else being equal, depth of field is inversely proportional to sensor pitch.
If you want an interviewee to be in sharp focus but the background to be soft, then a 1/3-inch format camera needs the background to be twice as far away as does a 2/3-inch format camera. Similarly, it is twice as hard to “pull focus” from one character to another to focus a viewer’s attention.
Fortunately, the new camcorders are HDTV—or are they? When light passes through an aperture, it is diffracted. The diffraction means a point of light will actually appear as a disk.
If the disk radius is smaller than the sensor pitch, then maximum resolution isn’t limited by diffraction. Sharpness, however (which is a function of both resolution and contrast ratio), is diffraction limited unless the disk diameter (not radius) is smaller than the sensor pitch.
The diffraction-disk radius or diameter is a function of the wavelength of light and the numerical f-stop. The sensor pitch is a function of the imager size and the number of sensor rows. For red light (at a wavelength of 630 nm), a 2/3-inch format 16:9 aspect ratio sensor, and standard definition, diffraction limits resolution at about f/14 and sharpness at about f/7. Most shooting will probably be at wider apertures and, therefore, not diffraction limited.
In a camera equipped with 1/3-inch format imaging chips with 1080 rows of sensors, red-light resolution will be diffraction limited at about f/3.5, and sharpness will be diffraction limited even with an aperture wider than f/1.8.
Virtually all shooting will be affected by diffraction.
None of this means that small-format HDTV camcorders are a bad idea. They’re not. In fact, their increased depth of field can make them ideal for well-lit news and surveillance purposes. But sensor-pitch issues do mean that a 1/3-inch format HDTV camcorder is not necessarily the ideal tool for all purposes.
Panavision’s Genesis digital-cinematography camera, which, with its HDCAM SR recorder may be considered an HDTV camcorder, uses an imaging chip with 21 times the area of a 1/3-inch chip. Unfortunately, it costs a lot more, too.
In other words, bigger means beggar.
Mark Schubin is an engineering consultant with a diverse range of clients, from the Metropolitan Opera to Sesame Workshop.
CATEGORIES
