2/24/2008 9T6WP
Preliminary Hardware Data Module
BCM7405
06/29/07
Functional Description
Bro a d c o m Co rp o r a ti o n
Document
7405-1HDM00-R
Advanced Video Decoder Page 1-29
•
BSP can perform external memory data validation. For example, BSP can verify the signature of the codes stored in, for
example, the off-chip program memory before the host CPU is authorized to execute these codes.
•
Authentication process: Use of challenge-response mechanisms to activate various busses or test ports (PCI, EBI, and
JTAG)
A
DVANCED
V
IDEO
D
ECODER
O
VERVIEW
The Advanced Video Decoder (AVD) module is a high-definition AVC/MPEG-2/VC-1/DivX/MPEG-4 P2 video decoder core.
The AVD module retrieves elementary stream video data placed into SDRAM by the data transport module, decodes the
video, and writes the decoded pictures back to SDRAM to be retrieved by video feeder in the video display subsystem.
The AVD core is capable of decoding one or more encoded elementary streams. The processing of such a stream has two
major components: front-end processing (the conversion of the code stream into fundamental components—motion vectors,
transform coefficients and the like) and back-end processing (actual generation and manipulation of pixels).
FGT block average logic is optional, it is available to compute block averages as an assist to the downstream FGT logic.
When enabled, this block monitors decoder pixel output and uses this data to calculate 8x8 block averages, which are written
to main SDRAM memory.
S
UPPORTED
P
ROTOCOLS
, P
ROFILES
,
AND
L
EVELS
The AVD module can decode the following code streams:
•
H.264/AVC main and high profile to level 4.1
•
VC-1 advanced profile @level 3
•
VC-1 simple and main profile
•
MPEG-2
•
MPEG still-picture decode
•
MPEG-4 Part 2
•
DivX 3.11, 4.11, 5.X, 6.X
The advanced video decoder module supports tools added in the AVC Fidelity Range Extensions (FRExt) amendment,
specifically 8x8 transform and Spatial Prediction modes, and adaptive quantization matrix required for High Profile support.
Additional capabilities include the following:
•
Error concealment
•
Multiple-stream support for up to sixteen low-resolution streams