Image Processing Reference
Smoothing or softening the image with a little blurring might remove some noise artifacts
and lead to better coding. But that may reduce the bit rate a little, since the coding of sharp
edges requires high-frequency coefficients to be included in the coded output. It is impor-
tant to understand how edges work in digital imaging before we can make informed deci-
sions about how to optimize them for compression.
If you are working on edges, apply the correction just to the luma channel. The
chroma channels can remain untreated unless the distortion is gross.
The slew rate is the capacity of your input circuits to adjust to a sudden and drastic change
in the intensity levels. If the slew rate is insufficient, the edges will become blurred. Figure
34-10 shows two slew-rate scenarios; one depicts going from black to white and the other
from gray to white. The slewing distance may be different, but because of the edge, the
slewing rate is just as fast because of the sharp transition. It is only the excursion that is
A poor slew rate will look as if the picture is blurred in the horizontal direction and
fine detail may be lost altogether. Thin vertical lines, for example, just disappear alto-
gether. This is evidenced in the example waveforms shown in Figure 34-11.
Applying additional filtering may control the sharpness of the detail in the video. But you
need to be careful to avoid overshooting. There are several artifacts that are undesirable,
and the wrong sharpness-filter settings will actually make your compression less effective.
Figure 34-10 Slew rate.