DPCM Predictors for NTSC color composite TV signals using phase adjustment of sampling

Claim 2

In a method of predicting a color television signal having a color carrier frequency f sc , of the type wherein consecutive lines of said signal are sampled at a frequency in excess of the Nyquist rate, the sample values are stored and a plurality of stored sample values are algebraically combined to obtain a predicted value of a current sample, the improvement comprising: sampling said television signal at a rate of (n/m) f sc , where n and m are integers, and n is an even number; and phase-shifting the sampling times of each line of said signal relative to the sampling times of the prior line of said signal.

Claim 3

The method of predicting a color television signal according to claim 1 wherein said sampling times are phase-shifted by an amount which is an integral multiple of (1/n) times the color subcarrier period.

Claim 4

The method of predicting a color television signal according to claim 2 wherein said sampling times are phase-shifted in order to obtain a horizontal distance of one-half the subcarrier period between samples on consecutive lines having the same subcarrier phase.

Claim 5

The method of predicting a color television signal according to claim 3 wherein said plurality of stored values comprises samples values B and C from said prior line of said signal and sample value A from the current line of said signal, and wherein the prediction X of the current sample value X is calculated according to the equation X = B + 1/2(A-C) where samples X and A have the same subcarrier phase as samples B and C, respectively.

Claim 6

The method of predicting a color television signal according to claim 4 wherein said color television signal is a NTSC color composite signal.

Claim 7

The method of predicting a color television signal according to claim 2 wherein said sampling times are phase-shifted in order to vertically align samples from consecutive lines.

Claim 8

The method of predicting a color television signal according to claim 6 wherein said plurality of stored values comprises sample values B and C from said prior line of said signal and sample value A from the current line of said signal, and wherein the prediction X of the current sample value X is calculated according to the equation X = A + C - B where samples X and A are vertically aligned with samples B and C, respectively.

Claim 9

The method of predicting a color television signal according to claim 7 wherein said color television signal is a NTSC color composite signal.

Claim 10

The method of predicting a color television signal according to claim 2 wherein said sampling times are defined by sampling pulses from a timing means, said timing means comprising: master clock means phase-locked to f sc for providing master clock pulses at a rate of nf sc ; a "divide by m" counter connected to receive said master clock pulses and providing sampling pulses at a rate of (n/m)f sc ; and decoding means, enabled once during each horizontal blanking period, for providing a preset pulse at the beginning of the current line of said signal when a count of k is detected in said divide by m counter, whereby said preset pulse simultaneously disables said decoder and resets said divide by m counter to zero so that the sampling pulses on said current line, which occur thereafter at every m th master clock pulse, are phase-shifted by k master clock pulses relative to the sampling pulses of the prior line.

Claim 11

The method of predicting a color television signal according to claim 2 wherein said sampling times are defined by sampling pulses from a timing means, said timing means comprising: master clock means phase-locked to f sc for providing master clock pulses at a rate of nf sc ; counting means connected to receive said master clock pulses; a decoder providing a sampling pulse each time a count of I is reached in said counting means first preset means for presetting said counter to a value of (-J) when a count of I is reached in said counting means, where m = (I + J), thereby providing "divide by m" operation; and second preset means for presetting said counter to a value of (-J-k) once during each horizontal blanking period when a count of I is reached in said counter so that the next sampling pulse will occur after (m + k) master clock pulses and the sampling times on said current line, which occur thereafter at every m th master clock pulse, are phase-shifted by k master clock pulses relative to the sampling pulses of said prior line.