Image Processing Reference
With the plethora of aspect ratios and built-in zoom and stretch modes on modern
TV sets, it is very likely that almost no TV services and movies are broadcast in the
aspect ratio that the program maker or movie director intended. But the human eye
and vision system is remarkably good at compensating for this and consumers seem
to not be unduly concerned. But viewers are occasionally surprised when meeting a
celebrity face-to-face to find that the person is not in fact as fat or thin as he or she
appears to be on TV.
Aspect Ratio Conversion
Compensation for aspect ratio is sometimes necessary when you compress the video.
Broadcasters use a device called an aspect ratio converter (ARC), and while this is very
convenient, it is possible for material to be converted backward and forward several times.
The result is a rather low-quality picture that appears to be over-cropped. This is because
the top and bottom are trimmed when going from 4:3 to 16:9, and the sides are trimmed
when going back to 4:3. Figure 5-19 shows the commonly used ARC output settings when
cropping the input to fit.
An ARC will also pad an image rather than cropping it. Figure 5-20 shows some
commonly used padding setups.
The starting point is essentially an analog representation since the world at large is ana-
log. It is very convenient to operate on video using digital techniques but in the end it is
presented to a Mark 1 eyeball, which operates in an analog fashion. It is likely to remain
that way unless the Borg collective from Star Trek assimilates us all. While we wait until
that happens, let's examine the conversion process step by step.
Analog Video Source Format
A video signal is a continuously varying light intensity that is detected by the imaging
electronics in the camera. Early cameras were monochrome (black and white) so there was
just the one signal.
Later on, color cameras were developed so there were three signals to be processed
and delivered. A color picture is composed of three monochrome rasters. The primary col-
ors, red, green, and blue, are chosen because the display is composed of three light
sources. Therefore additional color mixing is necessary. So green plus red equals yellow.
Color information is reviewed again in more detail later in this chapter.
Video is most often represented as a black and white image (known as luma or Y')
with some color components (known as chroma or C B and C R ) described separately. This
is quite a different and complex way to describe a picture compared with an RGB signal,
which is very simple. The luma plus chroma representation is actually more compact and
requires less information to describe it.