GB2181027A - Digital to vestigial sideband conversion - Google Patents

Abstract

Television and other video information is usually handled in vestigial side band (VSB) analogue form, but a digital form is better for transmission and storage. Hence the VSB signal is placed in a delta sigma (pulse density modulated) form in a coder. At the receiving end the digital signal is applied with a local clock to a decision circuit 1 whose output is applied to a gate 8, also clock controlled. The parameters are such that the gate output is a carrier at the clock frequency amplitude modulated by the TV or video. This is applied to a band-pass filter whose output is the original VSB signal. <IMAGE>

Description

SPECIFICATION Digital to Vestigal Sideband Conversion This invention relates to circuit arrangements for conversion between information represented in digital form and in vestigal sideband form.

Video signals as handled in present day networks and broadcasting systems are conveyed in analogue vestigal sideband (VSB) format. Current equipment, such as television receivers, is equipped to receive such a signal format and not a digital format. Digital encoding is however to be preferred to an analogue format since it achieves a constant quality which is unaffected by transmission, switching or storage as long as no errors are introduced.

Thus a digital encoding method whose signals are easily converted to VSB would encourage the use of digital transmission, switching and storage, with conversion to VSB at the point at which the signal has to enter an existing VSB system or apparatus. A digital code which is readily converted to a baseboard composite analogue video signal, e.g. PAL, is described in "An optical 1-bits video link", by E.

Roza, Proc 7th ECOC, 1981.

The principle of this will now be described with reference to Fig. 1. The digital code is regenerated by a decision circuit 1, in which it is, in effect, integrated into a clock pulse train. It is then low-pass filtered by filter 2. The digital code is such that the output of the filter is an accurate replica of the original composite (PAL) video signal. It is then amplified by an amplifier 3 to obtain the desired signal level.

In the coder, the digital version of the signal is subtracted from the original input signal in a differencing amplifier 4. The low pass filter formed by resistor 5 and capacitor 6 may be included to perform a similar function to the filter 2. In any case, the succeeding amplifier/encoding device 7 generally includes several stages, and contains frequency-shaping circuits such as integrators, e.g.

as shown at Fig. 10 of a paper "A Use of Double Integration in Sigma Delta Modulation", by James C. Candy, IEEE Transactions on Communications, Vol. Comm. 33, No. 3, March 1985 at p 249 et seq.

In delta sigma modulation, the AC signal, in this case video, to be digitized is sampled at a high rate and each sampling causes a 1 bit to be sent if the signal amplitude has increased since the last sampling and a 0 bit if it has decreased. A direct current condition is sent as an alternation of 1 and 0 bits.

The actual conversion into ths delta sigma form is effected in the coder 7, which receives as its inputs a clock signal and the output of the differencing amplifier 4.

An object of the invention is to provide an improved coding/decoding method of the same general type as that described above.

According to the invention, there is provided a decoder for decoding from a delta sigma bit stream conveying information into a vestigial sideband analogue format, which includes a two-input decision circuit to one input of which the bit stream to be decoded is to be applied and to the other input of which a clock signal is applied, a gate to which the clock pulse signal and the output of the decision circuit are applied, the arrangement of the decision circuit and of the gate being such that the output of the gate is a train of return to zero pulses whose modulation includes a carrier at the clock pulse bit rate or a multiple thereof and upper and lower sidebands which correspond to the said alternating current information, and a bandpass filter to which the output of the gate is applied, which bandpass filter passes one of the sidebands plus the carrier to produce a vestigial sideband analogue signal which reproduces the said information.

Embodiments of the invention will now be described with reference to the accompanying drawings, in which: Fig. 2 shows a decoding arrangement embodying the invention; Figs. 3 and 4 are two alternative versions of the differing amplifier 4, Fig. 1; Fig. 5 is a graph useful in explaining the invention.

The decoding arrangement shown in Fig. 2 is an improvement of that shown in the right-hand side of Fig. 1. In Fig. 2 a gate 8, e.g. an OR gate, although NOR, AND or NAND gates are also usable, is inserted into the decoder at the point indicated by the cross in Fig. 1. This gate, in response to the incoming bit stream from the decision circuit 1 produces half-width return-to-zero (RZ) pulses, each of about half a bit time wide. When an output such as produced by the gate 8 is examined in the frequency instead of the time domain, it is found to contain a carrier at the clock frequency and upper and lower amplitude modulation sidebands. Thus in one case where the clock frequency was 100 M bit/ sec. these were two 3 MHz sine waves, one on each side of the carrier.This double sideband signal is then filtered by a bandpass filter 2a, thus producing a VSB signal whose carrier is at the clock frequency.

Fig. 2 also includes a frequency charger which consists of a mixer 9, with its local oscillator 10 and a bandpass filter 11. These components are included when it is necessary to translate the VSB signal to a different frequency allocation. These components are conventional. The filter 11 feeds the output amplifier 3a, which is a radio frequency amplifier used to distribute the VSB signal.

At this point the process of conversion between the digital signal and the amplitude modulated input to the filter 2a will be considered.

The delta sigma modulation technique used, with its relatively high sampling rate is in effect a form of pulse density modulation (PDM). With such modulations if a train of negative-going pulses at the maximum sampling rate is applied to a low pass filter all that remains is the average level of the pulses. However, with a bandpass filter, since pulses are present all the time, a strong carrier of the same amplitude (approximately) as the pulses is produced, and this appears as a constant signal envelope.

The next condition to be considered is where we are coding the maximum positive level it can handle, applied as a DC condition. No negative going RZ pulses are sent, and when this is passed through a low-pass filter, the resulting output is a steady state of the maximum voltage level. Using a bandpass filter, since there are no pulses, there is no carrier output.

We now consider a digital bit stream in which alternate negative going pulses are omitted, i.e. we have half the pulse rate, with pulse width unaltered.

At the output of the low-pass filter the signal is three-quarters of the way between the maximum negative-going value and the maximum positive value. In the case of the bandpass filter, we have pulses half the time at the filter's input, and this filter has a small percentage band-width. This has two effects, (a) the filter does not respond with the same amplitude as the pulses driving it, and (b) the filter "rings", filling in the time between the exciting pulses with a sine wave. The net result of this is to produce a carrier of half the amplitude of that produced in the first case.

We now consider the case when the encoded waveform applied to the filter is a TV waveform. In this case the negative-going pulses occur more or less frequently dependent on the nature of the waveform, so the output of the low-pass filter follows the encoded waveform. Note that the encoded waveform changes occur more slowiy than the bit rate of the PDM pulses. At the most negative level, which is usually the sync. tip, the saturation corresponds to the maximum PDM pulse rate, and at the most positive (peak white) level to that due to the maximum positive level discussed above.

Intermediate levels contain an intermediate rate of pulses. Now Fig. 2 is the bandpass filter case with an encoded TV waveform applied to it. Here the level of the carrier varies with the encoded TV signal, i.e. it is amplitude modulated therewith. Thus the sync-tip corresponds to the maximum PDM pulse rate, so carrier amplitude is greatest. At peak white, i.e.

maximum continuous amplitude, the carrier can disappear completely or almost compietely. In the composite signal case we get an output formed by an amplitude-modulated carrier, which accurately reproduces the TV signal from which the PDM bit stream was derived. In this case the carrier is almost extinguished at peak white levels.

The above discussion of what occurs when using a low pass filter (as in the prior art) and a bandpass filter (as provided by the present invention) is expressed in terms of negative-going pulses because the present embodiment uses negativegoing pulses.

In another embodiment of the invention, the width of the RZ pulses is made small relative to a bit period, instead of half the bit period, so that the results of the decoding initiate harmonics of the clock frequency. These harmonics are also amplitude modulated by the desired signal, in which case the frequency changing components 9--100-11 are unnecessary.

An apparatus for receiving a VSB signal can in general also receive an amplitude-modulated signal. In this case the band-pass filter may be omitted. It is only then needed where the bandwidth of the amplitude-modulated signal and the interference of the digital code will cause interference with other signals in different frequency allocations at the VSB receiving equipment.

The differencing amplifier 4, Fig. 1, may have the detailed structure shown in Fig. 3 to shape its frequency response. Here the two inputs, the TV signal and the digital form of that signal are applied to the two inputs of a first amplifier stage 12 which feeds a second amplifier stage 13, across which is connected a CR circuit. The values of these components, and of the other resistors and the capacitor are such as to give the noise spectrum the desired frequency response.

When the composite video signal includes a chrominance band, the amplifier 4 is as shown in Fig. 4. This includes in the circuit across the said amplifier 13 a network 14 which is resonant in the chrominance band, which increases the gain of the amplifier 4 in the chrominance band.

It is well-known that increasing the gain of the amplifier 4 reduces the noise by a corresponding amount, the limit to the amount of gain which can be applied being fixed by the classical theory of feedback systems. The implantation shown in Fig. 4 permits some increases in gain in the chrominance band, which decreases chrominance coding noise.

This is at the expense of being able to apply rather less feedback at other frequencies.

The effect on the spectrum of coding noise observed at the output of the filter 2, or in the VSB case band-pass filter 2a, is shown in Fig. 5 where N is the noise distribution for an amplifier as shown in Fig. 3, while Na is that for an amplifier as in Fig. 4.

The curve S represents the signal amplitude.

Claims (5)

1. A decoder for decoding from a delta sigma bit stream conveying information into a vestigial sideband analogue format, which includes a twoinput decision circuit to one input of which the bit stream to be decoded is to be applied and to the other input of which a clock signal is applied, a gate to which the clock pulse signal and the output of the decision circuit are applied, the arrangement of the decision circuit and of the gate being such that the output of the gate is a train of return to zero pulses whose modulation includes a carrier at the clock pulse bit rate or a multiple thereof and upper and lower sidebands which correspond to the said alternating current information, and a bandpass filter to which the output ofthe gate is applied, which bandpass filter passes one of the sidebands plus the carrier to produce a vestigial sideband analogue signal which reproduces the said information.

2. An information handling system, which includes a coder for converting alternating current information in vestigial side band analogue format into a delta-sigma bit stream, which bit stream is transmitted over a communications medium, and a decoder as claimed in claim 1 connected to said medium so as to receive a digital bit stream therefrom.

3. A system as claimed in claim 2, in which the coder includes an amplifier provided with a feedback circuit which includes a tuned circuit resonant in the chrominance band, whereby amplification and noise rejection in the chrominance band are improved.

4. A system as claimed in claim 2 or 3, in which the output of the band-pass filter is applied to a mixer which also has an input from a local oscillator, and in which the output from the mixer is applied via a further band-pass filter to an amplifier whose output forms the system's analogue vestigial sideband output at a frequency different from the digit rate of the system.

5. A decoder for decoding from a delta-sigma bit stream conveying video information into a vestigial side band format, substantially as described with reference to the accompanying drawings.