Abstract :
The results of experiments with a coder for compressing the bandwidth of television signals are reported. A complete television system was used consisting of a transmitter, a receiver and two channels. At the transmitter a ?slope-feedback? coder reduced the bandwidth of the television signal before transmission; waveforms specifying scanning-spot brightness and scanning-spot position were separately conveyed to the receiver; the two waveforms were used to reconstruct the television picture at the receiver. It is shown that, under virtually noiseless conditions, the total bandwidth for the television signal can be reduced some 3-7 times below the normal one. For this extreme compression ratio, some distortion is evident in the received picture even when slope-feedback coding is used. This mainly takes the form of incorrect shading of large surfaces and some defocusing where white dots are excessively reproduced; some spatial distortion is also evident. These detrimental effects could be reduced with improved experimental techniques. The experiments show quite definitely that cyclic patterns which were suppressed by narrow channel bandwidth can be restored by slope-feedback coding using the same bandwidth. Some features of the method are (a) the irregular scanning motions in the horizontal and vertical directions, (b) the variation of about 10% in the number of pictures per second (each picture contains the same number of scanning lines), (c) that synchronizing pulses are not needed, and (d) that relatively standard apparatus is required, although the cathode-ray tube at the receiver must have a wide contrast range. The lowest signal/noise ratio which can be tolerated in each of the channels does not appear to be excessive; for one channel it is much the same as for normal television (30dB or so), and for the other it is higher, i.e. 45dB. It must be emphasized that the experiments are limited to indicating some of the potential features of the two-channel method; the- y do not constitute a crucial test of the method whose final value must be assessed with more refined experimental techniques than are available to the author.
