High Tech Times- What Video Really Means

High-Tech Times Article 016

What Video Really Means

Last month, I talked about the forthcoming High-Definition Television (HDTV) standard, but we still need to consider that our current NTSC (National Television Standards Committee) TVS represent a monumental legacy that will be with us for many years to come. So here’s a bit of history, as well as how this older technology works.

By the early 1950s, the color TV transmission system was ready to be rolled out, with the requirement that it be compatible with the million or so monochrome TVS already in homes. Basically, if a TV station used three monochrome cameras, each with one of the three primary color filters (Red/Green/Blue) in front of its lens, a color image could be captured and transmitted to one of the new tri-color TVS. This was straightforward, but totally unacceptable, as it would use three full-bandwidth TV channels for one full-color program.

So the video engineers had to come up with a viable alternative to crunch both color and monochrome detail info into one TV channel. Looking at the chart, you can see that a color TV camera does indeed have three full-bandwidth image-sensors, each “seeing” only one primary color. But the smart video engineers could see that the data from each sensor is identical, which allowed them to create a scheme that would eliminate redundant info. The duplicated data are the brightness of the scene and the associated detail, both of which are seen in a monochrome image. In other words, if we add together the images from each sensor, we obtain a monochrome image having brightness and maximum detail.

In our NTSC TV system, we refer to this composite image as “luminance” or “Y information.” In the camera, this Y signal becomes a reference for determining what image information is unique, and which info can thus be eliminated to conserve bandwidth. Now if we compare the Y info from each of the three color sensors by digital subtraction, we get a signal that represents the unique data provided by that sensor. Conveniently, this process works just like the human eye that also has two types of image-sensors: rods and cones. Cones are receptors for color info; in dim light, you see details first and color second.

Looking at the chart, the RGB components are full bandwidth, and are modified by gamma correction. Gamma correction is the “normalizing” of each sensor’s signal so that the picture from each has the appearance of a linear brightness, which is needed because your TV’s cathode-ray tube (CRT, the big tube that you watch) light output is not linear with respect to the supplied electrical input.

The matrix decoder performs the math on the RGB signals to create the luminance (Y) reference signal, and the two channels of color information, which are also called the difference signals. So taking Red minus luminance (R-Y) and Blue minus luminance (B-Y), we now have three unique color data sets. But if you remember your high school algebra, if we have two sets, we can calculate the third; so we can discard one channel of color info and derive it later as long as we have the two color-difference signals. As the color-difference signals are lower-bandwidth than Y, we use them to transmit the color TV images that you all know and love.

You’ll notice that the chart references “I” and “Q” signals from the color encoder. These signals are still another way to decrease bandwidth. The I component is the “in-phase” signal that is controlled by the R-Y channel, while the Q component is phase-shifted 90 degrees; when the I and Q signals are added back to the Y luminance, our standard composite NTSC signal is the result.

Now let’s take a look at the TV image itself. As technology back in the 1950s wasn’t quite as sophisticated as today’s, the video engineers decided to use a rather simple method to squeeze everything into the 6 MHZ of bandwidth that was allotted for each station. Thus, TV uses two “interlaced” images every 1/30 second; these two images are called “fields,” and together they make up one “frame” of video. Our system sends one-half the picture in one field, and 1/60 second later sends the other half in the second field.

This interlacing makes today’s TV signal much less clear and precise than the image on your computer monitor, as there is a built-in “flicker.” This flicker would be immediately noticeable except for the human eye’s tendency to treat a series of images of 20 still-frames per second as continuous motion. As your TV shows 30 frames each second, and even your local theater shows them at 24 frames per second, your eyes and brain “fool” you into thinking that you are perceiving a continuous stream of images.

But you can also immediately tell the difference when you see one of the new high-definition TV (HDTV) screens. Instead of interlacing images, nearly all HDTV screens use a “progressive” image-processing technique that is very similar to what you see on your computer monitor where the entire image is shown at once (there is one approved HDTV standard that would include interlaced images).

The HDTV images are also much higher resolution than your TV. The standard NTSC TV resolution is only 512X486 pixels (or picture-elements, which are the smallest parts of an on-screen image. The lowest approved HDTV resolution is 640X480, and most TV stations are giving consideration to 720X480 and 1280X720 resolutions, both of which are easier to work with and easier on the eyes than TV’s resolution.

Over the next few months, I plan to discuss how our local television stations are implementing HDTV, as well as what the differences will be to each of us, from even more stations to multicasting to interactive TV. See you next month.