Imaging is Everything (page 2)

Having passed through the lens and iris, light then hits the image sensor. At the heart of the sensor assembly is a Charge Coupled Device (CCD) made up of hundreds of thousands of light-sensitive pixels arranged in a rectangular grid. These pixels store up an electrical charge in proportion to the amount and duration of light hitting them. Every 60th of a second, the camcorder reads these charges, combines them to create an image and discharges the CCD.

Without any engineering trickery, a CCD would respond only to the amount of light hitting it, not the color of the light. In other words, a CCD is by nature a color-blind device. Camcorder designers use two different approaches to extract color information from monochrome sensors, approaches that split the camcorder field into two camps.

Single-CCD camcorders are just what their name implies: models that use a single CCD sensor to handle all the image-making duties. Such a camcorder derives color information from the sensor by covering it with a mosaic of colored lenses. Like a microscopic stained-glass window, this "color mask" alternates red, green and blue (or cyan, yellow, magenta) panes over individual pixels. With some clever electronic processing, the camcorder can derive both a brightness (luminance) and color (chrominance) signal from the single CCD chip.

Three-chip camcorders use a trio of CCDs, each specializing in a certain color. By using a complex prism block or arrangement of mirrors and filters, a 3-CCD camcorder splits the light coming through the lens into three color components. Light from each of the color components (red, green and blue, for example) goes to its own sensor. The camcorder combines the output of these three chips to create a full-color video signal.

Single-CCD imaging systems offer the advantage of being smaller, lighter, less complex and cheaper to manufacturer. Three-CCD systems, though larger and costlier, generally deliver more accurate color at a higher resolution. Three-CCD designs may also deliver a hard-to-define improvement in image depth and realism. Three-CCD camcorders often have better lenses than their single-chip counterparts, a fact which also contributes to an improvement in video quality.

Recent years have seen a trend toward smaller and smaller image sensors in camcorders, from 1/2-inch to 1/3-inch to today's tiny 1/4-inch designs. A smaller CCD doesn't just mean a smaller sensor assembly; it means a proportionally smaller lens as well. Every aspect of a lens's design points back to how large an image it needs to create. If a lens needs to bathe a 1/4-inch sensor in light instead of a 1/2-inch sensor, designers can shrink the lens assembly considerably. This translates to smaller, cheaper and more compact camcorders.

Since a CCD's sensitivity is proportional to the surface area of each pixel, a smaller sensor will be less-sensitive to light if all other variables are held equal. In reality, the variables aren't held equal. CCD manufacturers have found ways to gather more light onto a smaller sensor. This gives today's smaller CCD designs low-light sensitivity on par with that of larger sensors.

Sensor resolution also plays a factor in image quality, up to a point. Once a sensor's resolution exceeds that of the recording system and tape format, there's little to be gained by increasing the sensor's pixel count. A 270,000-pixel CCD delivers ample resolution for a standard format like 8mm or VHS. Will a 470,000-pixel sensor result in sharper images in these formats? Probably not. Where extra pixels can be put to good use in with digital zoom and image stabilization.

Camcorder manufacturers have been trying to put a damper on the shakycam for many years now, introducing various image stabilization schemes that sense camera movements and compensate for them by shifting the image. Some systems use tiny motion sensors to detect motion, while others analyze the CCD image itself.

The simplest image stabilization is called electronic image stabilization (EIS). It shifts the CCD image to compensate for what it perceives to be unwanted movements. First, it crops the outer edges of the CCD image and expands the inner portion to fill the screen. It then slides this active image area around on the face of the CCD to compensate for camera movement. If a camera wiggle causes your subject to jump left 14 pixels on the CCD, the active image area also moves 14 pixels to compensate.

Because EIS throws away image pixels, some systems cause a noticeable drop in image resolution. Other EIS schemes use a higher-resolution CCD than the format really requires, allowing for lossless EIS.

The other method of quelling camera shake is optical image stabilization (OIS). OIS shifts elements in the lens itself to compensate for camera movement. This basically "steers" the light in the opposite direction of the camera shake. Because OIS performs its magic in the lens itself, there's no resolution loss whatsoever.

In the past, OIS systems have used a variable prism to steer the light. This prism is basically two pieces of glass connected with a flexible bellows and filled with optical fluid. Motors flex the bellows on two axes to change the path of the light. Manufacturers couldn't make the variable prism small enough for use in ultra-compact camcorders, so Canon recently came up with a different optical stabilization system.

Called an "optical lens shift" mechanism, this OIS scheme moves a group of elements on two axes to eliminate shake. This system is simpler than the variable prism, and can be made much smaller. Future compact camcorders will use optical lens shift technology.

There's a lot going on inside your camcorder to turn light into a usable video image. Some of this engineering wizardry is beyond your control, and some of it is yours to command. Armed with a better knowledge of what you can and can't control, it's time for you to start coaxing the best-possible images out of your camcorder.

• Digg This!
• del.icio.us
• Technorati
• StumbleUpon
• Reddit