20 ExtroNews 14.3 Fall 2003
Consider Figure 5, which shows the
second source from Figure 1 being scaled. In
this example the signal of a PC is delivered
at 1024 x 768 to an Extron ISS 408, then
to a projector at 1024 x 768. This example
represents one of the greatest challenges that
a scaler can meet head on, delivering one
pixel out for one pixel in, or as many pixels out
as in. It is accomplished by the scaler in this
example by sampling one pixel on the input
for each pixel output.
Clock, or pixel clock as it is sometimes
called, can be adjusted in a number of
different ways. In the ISS 408, adjusting the
horizontal size is done by changing the
number of samples taken across a given time
of active video, as shown in Figure 5. If the
number is increased the image gets smaller,
and vice-versa. By adjusting the size in this
way, the scaler is also adjusting the clock, or
the number of samples per pixel.
Figure 5 also demonstrates an additional
impact of adjusting the size in this way. The
lower image in each section of Figure 5
shows an oscilloscope readout of the number
of pixels sampled on the input versus the
number generated on the output. The top
image in each section shows the resulting
image, and in some cases the artifacts that
occur as a result of the clock speed. The test
pattern is a pixel on, pixel off pattern, where
every other pixel is turned on, and then off,
and can be generated in many ways with
computers or test generators.
Current technology allows for clock speeds
that sample at a rate of one sample per pixel.
Over sampling, or taking many more samples
than there are pixels, would be preferred,
however IC performance and cost currently
limit most scaler technology to 1:1 sampling
Optimizing Your Image with Auto-Image (continued)
Comentarios a estos manuales