SLES243E
– JULY 2009 – REVISED MARCH 2011
3.5
Scaler
Each video decoder has an independent horizontal and vertical scaler, which supports D1 to half-D1 or
CIF conversion.
Table 3-2 gives the details of video resolution including un-cropped and cropped.
Table 3-3 shows the video resolutions converted by the scaler.
Table 3-2. Standard Video Resolutions
Uncropped
Cropped
Format
NTSC
PAL
NTSC
PAL
D1
720 x 480
720 x 576
704 x 480
704 x 576
Half-D1
360 x 480
360 x 576
352 x 480
352 x 576
CIF
360 x 240
360 x 288
352 x 240
352 x 288
Table 3-3. Video Resolutions Converted by the Scaler
Active Output
Scaling Ratio
Format
Horizontal Scaling
Vertical Scaling
Total Pixel
Resolution
NTSC
1:1
858 x 525
720 x 480
D1
PAL
1:1
864 x 625
720 x 576
NTSC
2:1
1:1
429 x 525
360 x 480
D1 to Half-D1
PAL
2:1
1:1
432 x 625
360 x 576
NTSC
2:1
429 x 262
360 x 240
D1 to CIF
PAL
2:1
432 x 312
360 x 288
3.6
Noise Reduction
A video sequence shot under low light condition, which is typical of video surveillance applications, can
contain lots of noise. Human eyes are very sensitive to oscillating signals, the visual quality degenerates
significantly even when the noise level is small.
Each video decoder uses a TI proprietary spatial filter to reduce video noise. For each frame of image, the
video noise filter (VNF) produces an estimate of the Y/U/V noise. Based on the noise estimates, the
firmware adjusts the threshold for Y/U/V filtering. The filtered video shows improved video quality and
lower compression bit-rate. The firmware can also utilize the Y/U/V noise estimates to make decisions to
disable color if the video noise is determined to be too high. This "color killer" decision bit can be used to
control another module that implements the color killing function.
The Noise Reduction can be controlled using I2C registers from 50h to 5Fh. This module can also be set
to bypass mode by I2C register 5Dh (Bit 0).
3.7
Auto Contrast
The Auto Contrast (AC) module can adjust the picture brightness automatically or manually (user
programmable) for better image quality. The goal of AC processing is to make the dark area brighter and
high-light area dimmer. This makes it possible for the viewer to see details hidden in the shadows. It also
prevents loss of details in the washed-out high light area. The AC processing is mostly for video
surveillance applications.
For each frame of image, the auto contrast module collects the statistics of its Y (luminance) values. The
AC algorithm implemented in the firmware processes the statistics and generates a look-up table (LUT).
This LUT is used to map each incoming pixel Y value to an output pixel Y value for the next frame of
image. The LUT is updated during the blanking period between two frames.
The Auto Contrast Mode can be controlled by using I2C registers 0Fh. This module can also be set to
disable mode by I2C register 0Fh (Bit 1:0).
22
Functional Description
Copyright
2009–2011, Texas Instruments Incorporated