BNC 1420 - Video Microscope - Manual (Page 28)

26
6.1 Video compression
Video compression algorithms like MPEG are great for normal TV
recording, but may not work properly when used for scientific video
sequences, which often obey completely different statistics. MPEG video
compression is particularly bad for video sequences containing small,
rapidly moving objects, which is exactly the kind of imagery likely in
microfluidics experiments. Therefore, scopePROworks entirely with
uncompressed video sequences, making no assumptions about the nature
of the imagery.
6.2 Recording speed
A standard monochrome RS-170 video signal converts into a digital data
rate of 8.9 MBytes/s, which can easily be read into PC RAM memory in
real time. It is also within the capability of modern, fast computers to write
to hard disk in real time at this data rate, provided the computer is not
overloaded by simultaneously executing other disk or CPU intensive tasks.
Color NTSC video signals convert into a data rate of 26.4 MBytes/s,
which can readily be written to RAM memory in real time, but may be too
high to write to hard disk in real time. In that case, some frames scattered
throughout the video sequence are lost, resulting in a stored video
sequence with time intervals of 33.3 ms between most images, but with
66.6 ms, 99.9 ms or some other multiple of the base frame interval
between some individual images.
Such lost frames are called dropped frames. Video sequences with
dropped frames are not suitable for accurate time history analysis, since it
is difficult to know afterwards exactly where frames are missing.
6.3 Buffering
To avoid the problem of dropped frames and to enable pre-trigger
recording, scopePRO stores digitized video date in cyclic a RAM buffer
simultaneous with the display. This is illustrated in the figure below.