Sierra Video Ponderosa 3G Series Routing Switcher - Manual (Page 100)

SIERRA VIDEO
94
“F”: Field Delay
The command "F" is used to specify the delay between the time a crosspoint change
request is received by the router and the time the crosspoint switch actually occurs. It
must be followed by a number giving the number of video fields of delay desired. If the
number is smaller than the smallest delay that the router can handle, the smallest delay
is used instead. If it is larger than the largest delay the router can handle, the largest
delay is used instead. Note that this command does not cause a delay in command
processing, as the “D” command does.
To understand this command more fully, consider the way that router software will
typically handle a crosspoint command. The last character of the command string, the
final “!” (exclamation) character, is received somewhere in a particular video field, call it
video field 7. The router parses the command string and, for each crosspoint it contains,
it puts the crosspoint in a buffer that is marked to be delivered to the crosspoint hardware
on a particular video field. Suppose that previously, an “F5” command has been received.
Then crosspoint commands whose final “!” command string character was received on
video field 7 would be placed in a buffer that is marked to be delivered to the crosspoint
hardware at video field 13 (7+5+1=13).
To understand the reason for adding 1 in the previous sum, consider an “F0” command: it
would ask for output at the very next video field, field 8 in our case. So, it is necessary to
add the “F” argument plus 1 to the field number on which the crosspoint command is
received to get the field number at which the crosspoint will be output.
A typical router will have a minimum delay that is between 1 and 2 fields. Suppose a
crosspoint command is received just before a vertical field mark. The software may be
able to prepare the crosspoint data and send it to the hardware when that vertical field
mark occurs, but the hardware itself typically has a one-field delay in it, so the soonest
that such a crosspoint would switch would be one field (plus a little) from when it was
received. If the command were received towards the beginning of a field rather than the
end of a field, the delay would be closer to two fields. Industry parlance is to call this a
one-field delay, because only full fields of delay are counted.
The delayfields argument of this command takes into account the hardware delay. So, if
a router has a minimum delay, including the hardware delay, of one full field, as
described in the previous paragraph, then a delayfields value of 1 causes this minimum
delay to be used. A delayfields value of 0 will also cause this minimum delay, because
the router uses its minimum if a smaller value is specified. A delayfields value of 2,
however, will add one more field to the minimum possible delay. Thus, delayfields
specifies the number of full fields of delay between end-of-crosspoint-command-string-
received and crosspoint-switch-occurs. Note that the actual minimum value of delayfields
depends on the particular router model.
Routers typically have a limit to the number of crosspoint commands they can process in
one field. First, there is an inherent delay in sending the command to the router, but
beyond that, the router requires time to parse the command and buffer up the crosspoint
data, plus it requires time to deliver the buffered data to the hardware when the desired
video field arrives. Each individual router has documentation to describe its limitations on
how many crosspoints it can process in a given amount of time.
Larger values for delayfields give the router more time to process commands. Although
the long-term average number of crosspoints that can be processed per unit of time is
unchanged, a larger delayfields value can improve router performance during a short
burst of many crosspoint commands. For example, suppose a large number of crosspoint
commands is sent to the router in a single large command. If delayfields is small, the
router typically wonít have time to parse and process all these crosspoint commands and
place the data in the crosspoint delivery buffer before the target video field arrives. By