Real-time data compression of broadcast video signals

A differential pulse code modulation system for data compression of a digitized video signal (PIX) comprising:

• a first adder for receiving the PIX signal and producing a difference value DIF signal;
• a 13-level quantizer for processing the DIF signal to a quantization level QL.sub.N signal;
• a non-adaptive predictor for converting the QL.sub.N-1 signal to a non-adaptive predictor NAP signal and being connected from the output of the 13-level quantizer to the adder through an inverter;
• a second adder;
• a value quantizer connected between the output of the 13-level quantizer and said second adder;
• means connecting the output of said non-adaptive predictor to said second adder;
• a first-line delay means;
• a second-line delay means connected serially with said first-line display means;
• a four-pixel delay means;
• a third adder;
• a divide by two means;
• means connecting the output of said second adder to said third adder through said pixel delay means and through said line delay means to different inputs of said third adder;
• means for connecting an output of said third adder to an input of said divide by two circuit to produce a predicted value PV signal at an output of said divide by two circuit means for connecting the output of the divide by two circuit to the first adder and to the second adder;
• Huffman encoder means receiving a quantization level QL signal from said 13-level quantizer;
• first in, first out FIFO rate buffer means;
• variable length, parallel-to-serial converter means;
• means for connecting said FIFO means between said Huffman encoder and the converter means whereby the converter means outputs coded serial data and clock pulses.

A method for coding a video signal comprising:

• sampling the composite analog video signal at a rate four times the color subcarrier frequency (4.times.3.579545 MHz);
• selecting the fourth previous pixel from the same line and the same pixel from two lines previous in the same field;
• averaging the two-pixel values to produce the prediction value (PV) of the current pixel value;
• means for generating a non-adaptive predictor NAP signal including a default NAP signal based on prestored values;
• subtracting PV and NAP signals from the current pixel value PIX (PIX-PV-NAP) to yield a difference value DIF signal;
• quantizing the DIF value in levels in accordance with stored values yielding a multilevel quantization value QL signal;
• Huffman encoding the QL signal at each said levels;
• converting the Huffman encoded signal to serial data and clock pulses.

A method of coding a digitized video signal (PIX) comprising the steps of:

• generating a non-adaptive predictor NAP signal;
• generating a predicted value PV signal;
• inverting the non-adaptive predictor value NAP signal;
• inverting the predicted value PV signal;
• adding the PIX, the inverted NAP and the inverted PV signal using twos complement addition to produce a difference value DIF signal;
• directing the DIF signal to a non-linear, non-adaptive quantizer to produce an output signal;
• directing the output signal of the quantizer to a non-adaptive predictor which produces said NAP signal, to a multilevel Huffman encoder, and to a 1-pixel delay;
• directing an output signal from the one-pixel delay to the multi-level Huffman encoder;
• directing the output of the Huffman encoder to a shift register to produce encoded serial data and clock signals.

A method of coding a video signal comprising the steps of:

• sampling the video signal at a rate four times the color subcarrier frequency of 3.579545 MHz to create a digitized video PIX signal;
• converting the PIX signal to a difference value DIF signal;
• converting the DIF signal to quantization level QL.sub.N and QL.sub.N-1 signals having a plurality of levels;
• and
• encoding each level of the QL.sub.N and QL.sub.N-1 signals to yield serial data and clock pulse signals.

A data compression system for digitized video signals (PIX) comprising:

• means for converting the PIX signal to a DIF signal;
• means for converting the difference value DIF signal to quantization level QL.sub.N and QL.sub.N-1 signals;
• and
• multilevel Huffman encoding means for converting the QL.sub.N and QL.sub.N-1 signals to serial data and clock pulse signals;
• a unique word detect means for outputting a serial data signal, a PIX signal and an enable signal;
• means for transmitting said serial data and clock signals to said unique word detect means;
• means for operating on the signals outputted by the unique word detect means to produced PIX/QV (PIX for pixels 1-4 of each line, and QV for all remaining pixels of each line) and QL.sub.N-1 signals;
• means for operating on the QV and QL.sub.N-1 signals to produce an reconstructed pixel RP signal;
• and
• means for operating on said PIX and RP signals to produce a reconstructed, digitized video output.