Equipment Description
Page 4-4
Instruction Manual: evolution 5000 E5500 Mobile Contribution Encoder
ST.TM.E10033.3
The 656 video feeds into the Video Compression Module where the MPEG
video processing takes place. This processing produces I frames; this
mode of operation is known as intra-frame coding.
A feedback path is fed to the motion-compensated Predicted Frame and
Difference Signal Generation function which subtracts a predicted frame
from the current frame to produce a frame difference or prediction-error
value. This prediction-error value is coded instead of the complete frame
information, thus affording a considerable saving in bit-rate; this mode of
operation is known as inter-frame coding. The prediction-error
information is augmented by motion vector information which is also
coded and enables recovery of the current complete frame information by
a Decoder.
The I frames and prediction-error values are processed in blocks of 8x8
pixels which are first transformed by a Discrete Cosine Transform
(DCT).For typical images, the distribution of the coefficients is not uniform
due to spatial redundancy, and the energy tends to be concentrated into
the low frequency coefficients. The number of bits required for a block is
reduced by not transmitting the near-zero coefficients and by quantising
the remaining ones. Quantisation results in a reduction of the number of
possible values to be transmitted. These values are then entropy-coded
using a Variable Length Code (VLC). The VLC allocates codes which have
different lengths depending upon the probability with which they are
expected to occur.
4.2.3
Frame Sequence
Depending on the predicted frame production method, P or B frames are
produced for inter-coding.
The regular transmission of an I frame, at user selectable intervals
(usually every 12 frames), provides a regular reference from which
predicted frames can be generated, thereby ensuring that predictions do
not become wildly inaccurate.
The run of frames from the start of an I frame to the start of the next I
frame is known as a group of pictures (GOP). Each GOP is identified by a
unique header.
The Encoder GOP structure may be selected to be:
•
IP, for non-B frame operation (IPPPPPPPPPPI);
•
IBP, for operation with B frames (IBPBPBPBPBPI);
•
IBBP, for successive B frames (IBBPBBPBBPBBPI).
•
IBB, (IBBIBBIBBIBBIBBIBBI) - a professional editing standard;
•
IB, (IBIBIBIBIBIBIBIBIB) - a superior professional editing standard;
•
I Frame, (IIIIIIIIIIIIIII) - for professional editing and compression.
