1135
6437E–ATARM–23-Apr-13
SAM9M11
48.3.1.6
Post Processing Features
Table 48-6.
Postprocessing Features
Input data format
YCbCr 4:2:0 planar
YCbCr 4:2:2 (YUYV)
Input image source
Internal source: VDEC decoder
External source: e.g. a software decoder or camera interface
Output data format
YCbCr 4:2:0 semi-planar
YCbCr 4:2:2 (YUYV)
Fully configurable ARGB channel lengths and locations inside 32 bits, e.g.
ARGB 32-bit (8-8-8-8), RGB 16-bit (5-6-5), ARGB 16-bit (4-4-4-4)
Input image size
(combined mode)
48 x 48 to 4672 x 3504 (16 Mpixels)
Step size 16 pixels
Input image size (stand-
alone mode)
48 x 48 to 2048 x 2048 (4 Mpixels)
Step size 16 pixels
Output image size
16x16to1280x720
Horizontal step size 8
Vertical step size 2
Image up-scaling
Bicubic polynomial interpolation with a four-tap horizontal kernel and a
twotap vertical kernel
Arbitrary, non-integer scaling ratio separately for both dimensions
Maximum output width is 3x the input width (within the maximum output
image size limit)
Maximum output height is 3x the input height – 2 pixels (within the
maximum output image size limit)
Image down-scaling
Proprietary averaging filter
Arbitrary, non-integer scaling ratio separately for both dimensions
Unlimited down-scaling ratio (e.g. from 16Mpixel to QVGA)
YCbCr to RGB color
conversion
BT.601-5 compliant [12]
BT.709 compliant [13]
User definable conversion coefficient
Dithering
2x2 ordered dithering for 4-, 5- and 6-bit RGB channel precision
Conditional area-based deinterlace filtering
Programmable alpha
channel
Constant eight bit value can be set to the alpha channel of the 24-bit RGB
output data to control the transparency of the output picture. The resulting
32-bit ARGB data can be used as input data for later alpha blending.
Alpha blending
Output image can be alpha blended with two rectangular
areas.(1) YCbCr
semi-planar 4:2:0 PP output format is not supported when performing
alpha blending.
The supported overlay input formats are following:
- 8 bit alpha value + YCbCr 4:4:4, big endian channel order being A-Y-Cb-
Cr, 8 bits each
- 8 bit alpha value + 24 bit RGB, big endian channel order being A-R-G-B,
8 bits each
RGB image contrast
adjustment
Segmented linear
RGB image brightness
adjustment
Linear