42
TANDBERG
FieldView Administrator Guide
D14098.02 June 2008
show a single complete picture from the video.
P-Frames contain only the differences between the current and the previous
•
picture in the video and therefore rely on the previous picture for proper
decoding.
Since I-Frames encode a whole picture they are usually larger than P-Frames.
MPEG-4 Visual and Temporal Artifacts
Smooth video is the result of presenting successive pictures at regular
time intervals. If one or more pictures are dropped from the sequence of
presentation, it will be noticeable as a short pause. This is called a temporal
artifact. A temporal artifact may also be followed by a secondary effect if
the dropped picture was the reference frame for a P-Frame. In this case, the
P-Frame will not decode properly and will result in a visual artifact. Visual
artifacts appear as blocks of scrambled colors, sometimes referred to as macro
blocking. To avoid this effect the TANDBERG FieldView Device detects dropped
frames and compensates by inserting the last successful frame.
MPEG-4 Group Of Picture
A group of pictures (GOP) consists of an I-Frame as an initial reference frame
followed by zero or more P-Frames. A GOP size of 1 (all I-Frames and no
P-Frames) is valid but may result in poor image quality since the compression
engine must still meet the target bit rate setting and to do so the quality of the
individual frames must be reduced. If a frame is dropped when the GOP size is
1, a temporal artifact will be noticed but no visual artifacts will result because
there are not any inter-frame dependencies.
Increasing the GOP size results in better image quality because the P-Frames
are smaller than I-Frames and so the excess bit rate capacity may be used to
encode a higher quality I-Frame as a reference. The trade off is that a dropped
frame will result in not only a temporal artifact, but also a visual artifact
(unless the frame that is dropped was the last frame in the GOP). For example,
if the GOP size is 10 and an I-Frame is dropped, it will result in a temporal
artifact due to the dropped I-Frame. In addition, the nine P-Frames following
will not decode properly resulting in nine frames of visual artifacts. The visual
artifacts will continue until the next I-Frame is encountered.
Guidelines for Using Media Configurations
Once you’ve chosen a target and peak bit rate, you can adjust the quality of
the video stream using the following guidelines. Remember that the target and
peak should not exceed the available bandwidth of the current network path.
For a better quality image:
Reduce the fps.
•
Increase the GOP.
•
Reduce the image size.
•
For better motion capture increase the fps, reducing image size if necessary to
maintain the desired visual quality.
If there is packet loss (e.g. due to interference, channel contention, varying
BW) and you see macro blocking or pauses:
Decrease the GOP.
•
Use I-only (GOP=1) but keep in mind that if I-frames are lost, the motion
•
will be choppy.
Note:
In very high packet-loss environments it may be necessary to record
Chapter 6
