JPH0115230B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a predictive encoding device that predictively encodes an input television signal, compresses the band into a digital signal with a small number of bits, and outputs the signal. This invention relates to an improved device that can multiplex information.

At the output of the predictive encoding device that predictively encodes the input television signal and converts it into a band-compressed digital signal, other information signals such as audio information, data signals, error correction bits of predictive codes, and transmission line frame bits are added. There are cases where it is desired to multiplex the data. In this case, other information bits must be multiplexed without degrading the encoding characteristics of the television signal itself. In response to such requests, conventional methods have reduced the amount of information generated in the television signal by setting a fixed value for several samples in the blanking interval, thereby reducing the amount of time slots (surplus) used to send new information. I was trying to get information (bits of information). Although this method allows information to be transmitted without deteriorating the encoding characteristics of the image portion of the television signal, it is not possible to transmit a faithful waveform during several samples during the blanking period. Therefore,
It is not possible to test transmission characteristics (for example, noise test) using the waveform during this period.

Alternatively, it is possible to obtain a time slot for multiplexing other information by slightly lowering the sampling frequency of the predictive encoder, but in this case, the sampling frequency changes. Therefore, it is necessary to change the prediction function, which not only causes a major change in hardware design but also causes deterioration of high frequency characteristics.

The purpose of the present invention is to solve the above-mentioned conventional drawbacks and
While keeping the transmission speed and encoding image quality constant, the waveform during the blanking period is also transmitted as is, and other information can be added without significantly changing the hardware of the television signal encoding system. An object of the present invention is to provide a predictive coding device for television signals.

The encoding device of the present invention includes a difference circuit that outputs the difference between an analog or digital input television signal and the output of a prediction circuit described below, and an output signal that receives the output signal of the difference circuit and reduces the number of bits using a certain algorithm. For television signals, the device comprises a quantizer that outputs a signal, a prediction circuit that inputs the output of the quantizer and predicts and outputs an image signal at the next clock time, and a buffer memory that stores the output signal of the quantizer. In the predictive coding device, a specific quantizer operates in a specific period within the blanking interval and converts and outputs the output signal of the difference circuit to a number of bits smaller than the number of output bits of the quantizer, and an occupancy rate of the buffer memory. control the occupancy rate of the buffer memory to the predetermined value by increasing or decreasing the number of output samples of the specific quantizer and inputting the sample to the buffer memory according to the comparison result; The present invention is characterized by comprising a control section that outputs the number of samples, and a multiplexing circuit that multiplexes the output signal of the control section, the output signal of the buffer memory, and any other information signal.

Next, the present invention will be explained in detail with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention. That is, the input television signal 1 is sampled and quantized every sampling clock by the analog-to-digital converter 4, and converted into a digital signal and output. The sampling clock is generated by the pulse generator 11 using the output of the transmission line clock oscillator 13. The pulse generator 11 also receives the input television signal 1 and generates a frame signal F for each frame.
Outputs (12 bits).

The output of the analog-to-digital converter 4 is input to a difference circuit 5, and the output of the prediction circuit 8 is input to the other input of the difference circuit 5. Therefore, the difference between the input value and the predicted value is outputted from the difference circuit 5. The quantizer 6 is a 5-bit nonlinear quantization circuit, which reduces the number of bits of the output of the difference circuit 5 using a certain algorithm and outputs it as a 5-bit digital signal. The digital signal is stored in the buffer memory 9 via the switch 12 and is also input to the prediction circuit 8. The prediction circuit 8 predicts the next television signal based on the input signal. Prediction of television signals can be done using, for example, a high-order prediction function (see 1981 Electronic Communication National University No. 1054, Institute of Electronics and Communication Engineers Communication Systems Research Group Material CS73-44, Institute of Electronics and Communication Engineers Communication Systems Research Group Material CS74-63, etc.). can be used.
This prediction function also has good prediction accuracy for color components of color television signals. The above is almost the same as the conventional predictive encoding circuit.

In this embodiment, a 3-bit nonlinear quantization circuit (hereinafter referred to as a specific quantizer) 7 that operates during a specific period within a blanking section of an input television signal is provided, and the specific quantizer 7 and the quantizer The output of the device 6 is arbitrarily selected and outputted by the switch 12. The switch 12 performs a switching operation under the control of the control section 14, and outputs the 3-bit signal of the specific quantizer 7 for some of the specific periods during the blanking period. The number of samples selected by the specific quantizer 7 is controlled by the control unit 14 so that the occupancy rate of the buffer memory 9 is a constant value. The specific quantizer 7 has the same low-order quantization characteristics as the quantizer 6, but has a 3-bit output. During the blanking period, the input signal changes little or is regular (sine wave) except for some periods (near both ends of the horizontal synchronization signal), so the prediction result of the prediction circuit 8 is very good. Fit. Therefore, the output signal of the difference circuit 5 exhibits a very small value, and it is possible to transmit the waveform with sufficient fidelity even with 3-bit quantization. For example, during the blanking period of the input television signal shown in Figure 2a, the prediction results are very accurate in periods A, B, and C, excluding both ends (rising and falling parts) of the horizontal synchronizing signal. . In this embodiment, for simplicity, only period A is used, and during this period, the specific quantizer 7
is selected and the 3-bit output is input to the buffer memory 9. During other periods, the output of the quantizer 6 is selected and the 5-bit output is input to the buffer memory 9. Note that the specific quantizer 7
The number of samples S to be selected is determined by the buffer memory 9.
The control unit 14 controls the sample number S to be a constant value (for example, 1/2), and the control unit 14 sends the sample number S to the multiplexing circuit 10 as a digital signal. The multiplexing circuit 10 includes a pulse generator 1
Following the frame signal F outputted by No. 1, the number of samples S is then input to other signals (for example, the output digital audio information of the analog-to-digital converter 15 that inputs the audio signal 2 and converts it into a digital value) and the buffer memory 9. Multiplexes and outputs read signals.
Control of reading and writing data to and from the buffer memory 9 is performed by the control section 14.

In this embodiment, the horizontal synchronization signal of the television signal,
Since the sampling frequency is not phase-synchronized, the number of sampling points within one horizontal section is not constant, and writing control to the buffer memory 9 is required. The capacity of the buffer memory is set to about twice the information bits (frame synchronization signal F, output sample number S of a specific quantizer, and other information (eg, audio information)) added to the digital frame. The control unit 14 compares the occupancy rate of the buffer memory 9 with 1/2 at the timing of starting horizontal scanning shown in FIG. However, when it is smaller than 1/2, the number of samples S is decreased by 2 samples, and the number of samples S is controlled so that the occupancy rate of the buffer memory becomes close to 1/2. to the receiving side. On the other hand, the specific quantizer 7 outputs 3-bit DPCM data every clock, and accumulates S samples in the buffer memory, and then the 5-bit DPCM data output from the quantizer 6 is transferred to the buffer memory 9 every clock. (see Figure 2c). As described above, the multiplexing circuit 10 multiplexes and outputs the read data (3-bit DPCM and 5-bit DPCM) from the buffer memory 9 following the frame signal F, the number of samples S, audio information, etc. (see FIG. 2). (see d).

For example, if the sampling frequency _S is 680 _H ( _H is the horizontal scanning frequency), the digital data generated within one horizontal scanning section is 3400 bits (5 x 680), but the average is twenty one
Since the sample is 3 bits, 21×(5-3)
= 42 extra bits are created. If we allocate 12 bits of these surplus bits to the frame synchronization signal F and 5 bits to the number of samples S, the remaining
25 bits can be used to transmit other information (eg audio information). For example, two samples of audio information of 12 bits per sample can be sent. If 3-bit output is selected also in sections B and C, even more surplus bits can be used for other information.

In this example, it is assumed that an average of 42 surplus bits is generated, and the capacity of the buffer memory is approximately 42 bits.
×2 = 84 bits. FIG. 3 is a time chart showing changes in the occupancy rate in the buffer memory. In the initial operation of the buffer memory, first, when electricity is turned on, writing is performed until the occupation rate reaches 1/2, and then writing is stopped. Then, at the start of the horizontal scanning section, the 3 bits output from the specific quantizer 7
By writing the DPCM signal from time a to time b, the occupancy rate increases. During this time, the multiplexing circuit 10 sends out 42 bits of frame synchronization signal F, specific sample number S, audio information, etc. Subsequently, from time b, 5 bits are read from the buffer memory every 1 clock, and 3 bits are written into the buffer memory. Therefore, during this period, the occupancy rate of the buffer memory 9 decreases, and at time c when the transmission of the 3-bit DPCM signal ends, it becomes approximately 1/2.
Return to Thereafter, 5 bits are read out and written 5 bits at a time every clock, so the occupation rate hardly changes. At the start point a of the next frame, if the occupancy rate of the buffer memory 9 is less than 1/2, the control unit 14 decreases the number of samples S by 2 samples, and if the occupancy rate is greater than 1/2, the control unit 14 decreases the number of samples S. By increasing S, the occupancy rate is controlled to be approximately 1/2. By this, 1
It absorbs fluctuations in the number of sampling points during the horizontal scanning period. The capacity of the buffer memory is required to be at least as large as the number of information bits added within one scanning period.

In the above embodiment, the number of bits of the quantizer 6 is 5, and the number of bits of the specific quantizer 7 is 3, but the invention is not limited to this. The specific quantizer may have a sufficient number of bits to reproduce a simple waveform during the blanking period. Further, the blanking period to be used may be a vertical blanking interval instead of a horizontal blanking interval. However, in this case, the capacity of the buffer memory becomes large. In addition, the output of the prediction circuit is converted into an analog value, the difference with the input analog signal is calculated, and this difference signal (analog) is converted into a digital value and output by a quantization circuit composed of a nonlinear A/D converter. The present invention can also be applied to a hybrid type predictive coding device. In this case, the specific quantizer is of course constituted by a nonlinear A/D converter. Note that without providing a separate specific quantizer 7, a limiter function can be added to the quantizer 6 so that only a total of 3 bits, the sign bit and 2 lower digit bits, are output during a specific period (for example, period A). It's okay.

As described above, in the present invention, since the part in which the waveform changes within the blanking period is small or simple is configured to be output with a reduced number of bits using a specific quantizer, the number of bits is reduced. Other information can be multiplexed by the amount reduced. Since the image waveform portion is not changed in any way, the image quality does not deteriorate, and the input waveform can be sufficiently reproduced even in the blanking section.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention;
FIG. 2 is a diagram for explaining the operation of the above embodiment,
FIG. 3 is a time chart for explaining the operation of the buffer memory of the above embodiment. In the figure, 1...input television signal, 2...audio signal, 3...output terminal, 4...analog-digital converter, 5...difference circuit, 6...quantizer, 7...
... Specific quantizer, 8 ... Prediction circuit, 9 ... Buffer memory, 10 ... Multiplexing circuit, 11 ... Pulse generator, 12 ... Switching device, 13 ... Transmission line clock oscillator, 14 ... Control section , 15...Analog-digital converter.

Claims (1)

[Claims] 1. A difference circuit that outputs the difference between an analog or digital input television signal and the output of the prediction circuit described below, and a quantizer that receives the output signal of the difference circuit and outputs an output signal with the number of bits reduced by a certain algorithm. , a predictive coding device for television signals comprising: a prediction circuit that inputs the output of the quantizer and predicts and outputs a picture signal at the next clock time; and a buffer memory that stores the output signal of the quantizer; a specific quantizer that operates during a specific period within the ranking interval and converts and outputs the output signal of the difference circuit to a number of bits smaller than the number of output bits of the quantizer; and a value determined as the occupancy rate of the buffer memory. A control unit that controls the occupancy rate of the buffer memory to the predetermined value by increasing or decreasing the number of output samples of the specific quantizer according to the comparison result and inputting the result to the buffer memory, and outputs the number of samples. and a multiplexing circuit that multiplexes the output signal of the control section, the output signal of the buffer memory, and any other information signal.