JPS62266989A - Highly efficient encoder - Google Patents

Abstract

PURPOSE:To attain a high compression without producing the deterioration in a restored picture such as a block distortion by performing a variable length encoding by a non-linear characteristic matching to the visual characteristic of a human being. CONSTITUTION:A blocking circuit 2 converts an input digital TV signal into a signal succeeding evely block of an unit for encoding. A dynamic range detection circuit 3 detects a dynamic range DR and a minimum value MIN every block. A subtraction circuit 4 inputs picture element data PD and forms picture element data PDI in which the minimum value MIN is removed. A bit length decision circuit 6 decides the bit length Nb considering the visual characteristic of the human being correspondingly to the range DR. A quantization circuit 5 quantized the data PDI by the bit length Nb. A framing circuit 7 inputs an encoding code DT, the range DR and the minimum value MIN as addition codes to apply the processing for an error correction encoding to the code DT and the addition codes, adds a synchronizing signal to from transmission data and transmits to a trasnsmission path through an output terminal 8.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a high-efficiency encoding device that compresses the average number of bits per pixel of image data such as a digital television signal.

[Summary of the invention]

In this invention, in a high-efficiency encoding system applied when transmitting image data such as digital television signals, one screen is divided into a large number of two-dimensional or three-dimensional blocks, and each professional video By encoding with variable bit length that adapts to the dynamic range narrowed due to the correlation of pixels within a block, pixel data within a block can be encoded with a compressed bit length, which is reduced compared to the number of bits of the original data. It is possible to form transmission data with a specified number of bits. Furthermore, when setting the bit length according to the dynamic range, the bit length is set non-linearly so that the maximum distortion changes according to the dynamic range, rather than keeping the maximum distortion constant. According to this invention, the compression ratio can be increased without degrading the quality of the restored image on the receiving side.

[Conventional technology]

As a method for encoding television signals, several methods are known in which the average bit n per pixel or the sampling frequency is reduced for the purpose of narrowing the transmission band.

As an encoding method that lowers the sampling frequency, image data is thinned out to 2 by subsampling, and the subsampling point and the position of the subsampling point used during interpolation are indicated (i.e., which subsampling point is above, below or to the left or right of the interpolation point? A method has been proposed that transmits a flag (indicating whether point data is used).

One of the encoding methods that reduces the average number of bits per pixel is D PCM (differenti
alPCM) is known. D P CM focuses on the fact that the pixels of a television signal have a high correlation and the difference between adjacent pixels is small, and this difference signal is quantized and transmitted.

Another encoding method that reduces the average number of bits per pixel is to subdivide one field screen into small blocks and encode the average value and standard deviation for each block and one bit for each pixel. There is something that transmits the code.

An encoding method that attempts to reduce the sampling frequency using subsampling has a sampling frequency of 2, which may cause aliasing distortion.

DPCM has a problem in that errors propagate to subsequent decoding.

The method of encoding in units of blocks has the disadvantage that block distortion occurs at the boundaries between blocks.

In order to solve the above-mentioned problems, the applicant of the present application has proposed a method as described in Japanese Patent Application No. 59-266407.
The dynamic range defined by the maximum and minimum values of multiple pixels included in a two-dimensional block is determined, and with a variable bit length adapted to this dynamic range,
We have proposed a high-efficiency encoding device that performs encoding.

FIG. 11 is used to explain the encoding of the variable gutter length adapted to the dynamic range that has been proposed previously. The dynamic range is calculated for each two-dimensional block consisting of (4 lines x 4 pixels - 16 pixels), for example. Furthermore, the minimum level (minimum value) within the block is removed from the input pixel data, where one sample is 8 bits.

The pixel data from which this minimum value has been removed is quantized. This quantization is a process of converting pixel data from which the minimum value has been removed to a representative level. The maximum allowable value of quantization distortion (referred to as maximum distortion) that occurs during this quantization is set to a predetermined value, for example, 4.

FIG. 11A shows that the dynamic range DR is (maximum value MA
The case where the difference between X and the minimum value MIN) is 8 is shown. (DR=8
), the center level 4 is taken as the representative level LO, and (maximum distortion = 4).

That is, when (O≦DR≦8), the center level of the dynamic range is taken as the representative level, and there is no need to transmit quantized data. Therefore, the required bit length Nb is zero. On the receiving side 3, decoding is performed using the minimum value M I N of the block and the dynamic range DR, using the representative level LO as the restoration value.

FIG. 11B shows the case where (DR=17), the representative levels are determined as (LO=4) and (LL=13), respectively, and the maximum distortion E is 4. Since there are two representative levels LO and Ll, (Nb=1).

In the case of (9≦DR≦17), it is (Nb-1).

The narrower the dynamic range DR, the smaller the maximum distortion becomes.

FIG. 11C shows the case (DR=35), and the representative level is (LO=4) (LL=13) (L2=22) (L3
Since there are four representative levels LO-L3, each defined as (Nb=31) and (E=4), (Nb=2). In the case of (18≦DR≦3.5), (Nb=2).

In the case of (36≦DR≦71), eight representative levels (
LO-L7) is used. The 11th diagram is (DR=7
1), the representative level is (L 6 = 4) (
L 1 = 13) (L 2 = 22) (L 3 =
31) (L4=40) (L5=49) (L6=58)
(L7=67), respectively. 8 representative level L
In order to distinguish between O to L7, (Nb=3) is set.

(In the case of 72≦DR≦143, 16 representative levels (LO to L15) are used.
R=143), and the representative level is (L8=76).
) (L9=85) (L10=94) (Ll 1=103
)(L12=112)(L13=121)(Ll 4=
130) (Ll 5 = 139) (LO to L 7 are the same as the above values). 16 representative levels (LO
~L15), (Nb=4).

In the case of (144≦DR≦287), 32 representative levels (LO to L31) are used. Figure 11F is (
DR=287), and the representative level is (L16=
148) (L17=157) (L18=166) (L1
i=175)...(L27=247)(L28-
256) (L29=265) (L30=274) (L3
1=283) (LO to L15 are the same as the above values). In order to distinguish between the 32 representative levels (LO to L31), (Nb=5) is set. Actually, since the input pixel data is quantized with 8 bits, the maximum value of the dynamic range DR is 255, and the representative level (L
28 to L31) are not quantized.

Television signals within one block are horizontal.

Since there is a two-dimensional correlation in the vertical direction and a three-dimensional correlation in the time direction, the level of pixel data included in the same block varies only small in the stationary portion. Therefore, the minimum level M shared by pixel data within a block
Even if the dynamic range of the data DTI after EN is removed is quantized using a smaller number of quantization bits than the original number of quantization bits, almost no quantization distortion occurs. By reducing the number of quantization bits, the data transmission bandwidth can be made narrower than the original one.

[Problem that the invention seeks to solve]

In the above-mentioned encoding device adapted to a dynamic range with a variable bit length, the maximum allowable distortion E has been determined to be 4, for example. If the maximum distortion value is increased, the bit length Nb becomes smaller and the compression ratio can be increased. However, when the maximum distortion is increased, block distortion occurs.

Therefore, an object of the present invention is to provide a high-efficiency encoding device that can achieve higher compression rates without causing deterioration of restored images such as block distortion.

In this invention, when the bit length Nb is determined, the maximum distortion is not kept constant for the dynamic range OR;
The maximum distortion is changed using nonlinear characteristics that match human visual characteristics, and the bit length Nb is made smaller. Generally, when there is a sharp change in brightness level within a block, that is, when the dynamic range DR is large, a small change in brightness level is difficult to notice. Therefore, in the present invention, in a block with a large dynamic range DR, the maximum distortion value is increased and the bit length Nb is decreased.

[Means for solving problems]

The present invention provides a maximum value M A
A dynamic range detection circuit 3 that calculates a minimum value MIN of a plurality of pixel data and a dynamic range DR for each block, and a minimum value MI from a plurality of pixel data for each block.
A subtracting circuit 4 that removes N, determines the bit length Nb that nonlinearly corresponds to the da inamiso clean range DR for each block, and generates pixel data PD' from which the minimum value MIN has been removed.
I is supplied, and a quantization circuit 5 and a bit length determination circuit 6 perform quantization with the determined bit length Nb, and a frame that transmits information representing the dynamic range DR of each block and the output DT of the quantization circuit 5. This is a high-efficiency encoding device consisting of a coding circuit 7.

[Effect]

The compression rate can be increased by variable bit length encoding adapted to the dynamic range DR. In particular, in this invention, when the dynamic range DR is large, block distortion does not occur even if the maximum distortion is large, so the bit length is made shorter. Therefore, the compression ratio can be increased without causing block distortion.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. This invention is made in the following order of items.

a. Configuration of the transmitting side; Configuration of the receiving side C; Block and blocking circuit d; Dynamic range detection circuit e; quantization circuit f; Modification a; Configuration of the transmitting side FIG. This shows the overall configuration of the (recording side). For example, a digital television signal in which one sample is quantized to 8 bits is input to an input terminal indicated by 1. This digital television signal is supplied to the blocking circuit 2.

The blocking circuit 2 converts the input digital television signal into a continuous signal for each two-dimensional block, which is a unit of encoding. In this example, one block is (4
The size is (line x 4 pixels - 16 pixels). The output signal of the blocking circuit 2 is supplied to a dynamic range detection circuit 3 and a subtraction circuit 4. The dynamic range detection circuit 3 detects the dynamic range DR and minimum value MIN for each block. The pixel data PD from the blocking circuit 2 is supplied to the subtraction circuit 4, and the subtraction circuit 4 forms pixel data PDT from which the minimum value MIN has been removed.

Further, the detected dynamic range DR is supplied to the bit length determining circuit 6. The via length determining circuit 6 determines the number of quantization bits (bit length Nb) in correspondence with the dynamic range DR. In this case, the via length Nb is determined in consideration of human visual characteristics. That is, when the dynamic range DR is large, the maximum distortion E is increased. −
For example, the bit length determination circuit 6 determines the bit length Nb according to the dynamic range DR as follows.

This determined bit length Nb is supplied to the quantization circuit 5. The quantization circuit 5 is supplied with the pixel data PDI from the subtraction circuit 4 after the minimum value has been removed.

In the quantization circuit 5, the pixel data PD[ is quantized using the bit length Nb for each block described above.

The encoded code DT from this quantization circuit 5 is supplied to a framing circuit 7. The framing circuit 7 is supplied with a dynamic range DR (8 bits) and a minimum value M I N (8 bits) as additional codes for each block. The framing circuit 7 performs error correction encoding processing on the encoded code DT and the above-mentioned additional code, and also adds a synchronization signal.

Transmission data is obtained at the output terminal 8 of the framing circuit 7,
This transmission data is sent out to a transmission path such as a digital line.

As described above, the encoded code DT has a variable number of bits for each block, but the bit length of the pixel data of the block is uniquely determined from the dynamic range DR in the additional code. Therefore, although variable length codes are used, there is an advantage that there is no need to insert redundant codes indicating data divisions into the transmitted data.

b. Receiving side configuration FIG. 2 shows the configuration of the receiving (or reproducing) side.

The received data from the input terminal 11 is sent to the frame decomposition circuit 1.
2. The frame decomposition circuit 12 separates the encoded code DT from the additional codes DR and MIN, and also performs error correction processing. The encoded code DT is supplied to a decoding circuit 13, and the dynamic range DR is supplied to a bit length determining circuit 14.

In the bit length determining circuit 14, the bit length of each block is determined from the dynamic range DR in the same way as on the transmitting side. This bit length is supplied to the decoding circuit 13. The decoding circuit 13 performs processing opposite to that of the quantization circuit 5 on the transmitting side. That is, the data after removing the 8-bit minimum level is decoded to the representative level, and this data and the 8-bit minimum value M
IN is added by the adder circuit L5, and the original pixel data is decoded. The output data of the adder circuit 15 is supplied to the block decomposition circuit 16. The block decomposition circuit 16 is a circuit for converting decoded data in the order of blocks into the same order as the scanning of the television signal, contrary to the blocking circuit 2 on the transmitting side. A decoded television signal is obtained at the output terminal 17 of the block decomposition circuit 16.

C. Blocks and Blocking Circuits Blocks, which are units of encoding, will be explained with reference to FIG. In this example, by dividing the screen of one field, the image shown in Fig. 3 is obtained (4 lines x 4 pixels).
A large number of two-dimensional blocks are formed. In Figure 3,
Solid lines indicate odd field lines; Dashed lines indicate even field lines. Unlike this example, the present invention can also be applied to a three-dimensional block composed of four two-dimensional regions belonging to each of four frames, for example.

The blocking circuit 2 will be explained with reference to FIG. 4, FIG. 5, and FIG. W6. For the sake of explanation, it is assumed that the screen of one field has a configuration of (4 lines x 8 pixels) as shown in Figure 5, and this screen is divided into two vertically and horizontally as shown by the broken lines. A case will be described in which the image is divided into four in the direction and eight blocks (2 lines x 2 pixels) are formed.

In FIG. 4, as shown in FIG. 6, input data A consisting of four lines (Th, ~Tht) is supplied to an input terminal indicated by 21, and input data A is supplied to an input terminal indicated by 22.
A sampling clock B (FIG. 6B) synchronized with is supplied. Numbers (1 to 8) indicate sample data of line Th, and number (if-18) indicates line T.
Show the sample data of h, respectively, and indicate the numbers (21 to 28
) indicate the sample data of the line Th, respectively, and the numbers (31 to 38) indicate the sample data of the line Ths, respectively. Delay circuit 23 whose input data A has a delay amount of Th
and is supplied to the delay circuit 24 with a delay amount of 2Ts (Ts: sampling period! 1lI). Further, the sampling clock B is supplied to the divide-by-2 circuit 27.

The output signal C (FIG. 6C) of the delay circuit 24 is supplied to one input terminal of the switch circuits 25 and 26, respectively, and the output signal D (FIG. 6D) of the delay circuit 23 is supplied to the other input terminal of the switch circuits 25 and 26. are supplied to the input terminals of the respective input terminals. The switch circuit 25 is controlled by the output signal E (FIG. 6E) of the frequency divider 27, and the switch circuit 26 is controlled by the pulse signal E.
is controlled by a pulse signal inverted by an inverter 28. The switch circuits 25 and 26 alternately select the input signal (C or D) every 2Ts. switch circuit 2
The output signal F from switch circuit 26 is shown in FIG. 6F, and the output signal G from switch circuit 26 is shown in FIG. 6G.

The output signal F of the switch circuit 25 is connected to the first input terminal of the switch circuit 29 and the delay circuit 30 has a delay amount of 4Ts.
supplied to The output signal G of the switch circuit 26 is 2Ts
The signal is supplied to the delay circuit 31 having a delay amount of . The output signal H of the delay circuit 30 (H in FIG. 6) is supplied to the third input terminal of the switch circuit 29. Output signal I of delay circuit 31
(FIG. 6I) is supplied to the second input terminal of the switch circuit 29 and the delay circuit 32 having a delay amount of 4Ts.

The output signal J (FIG. 6J) of the delay circuit 32 is supplied to the fourth input terminal of the switch circuit 29.

The output signal of the frequency divider 27 is supplied to the frequency divider 33, and an output signal K (K in FIG. 6) is formed. The switch circuit 29 is controlled by this signal K, and the first . Second. The third and fourth input terminals are sequentially selected.

Therefore, the signal taken out from the switch circuit 29 to the output terminal 34 is as shown in (7) in FIG. In other words,
The field-by-field order of data is converted into a block-by-block order (for example, 1-2-11-12). Of course, the actual number of pixels in one field is much larger than in the example shown in FIG.
It is converted into the order of each block shown in the figure.

d. Dynamic range detection circuit FIG. 7 shows the configuration of an example of the dynamic range detection circuit 3. As described above, the image data of the area that needs to be encoded is sequentially supplied from the blocking circuit 2 to the input terminal 41 for each program. This input terminal 41
The pixel data from is supplied to a selection circuit 42 and a selection circuit 43. One selection circuit 42 selects and outputs the one with a higher level between the pixel data of the input digital television signal and the output data of the latch 44.

The other selection circuit 43 selects and outputs the one with a smaller level between the pixel data of the input digital television signal and the output data of the lunch 45.

The output data of the selection circuit 42 is supplied to the subtraction circuit 46 and is also taken into the latch 44 . The output data of the selection circuit 43 is supplied to the subtraction circuit 46 and the latch 48, and is also taken into the latch 45. Latches 44 and 45
A latch pulse is supplied from the control section 49. The control unit 49 is supplied from a terminal 50 with timing signals such as a sampling clock and a synchronization signal that are synchronized with the input digital television signal. The control unit 49 controls the latch 44
．． 45 and latches 47 and 48 at predetermined timing.

At the beginning of each block, the contents of latches 44 and 45 are initialized. The latch 44 is initialized with all “0” data, and the latch 45 is initialized with all “1” data. Among the pixel data of the same block that is sequentially supplied, the maximum level is the latch. 44. Also, among the pixel data of the same block that is sequentially supplied,
The minimum level is stored in latch 45.

When the maximum level and minimum level detection is completed for one block, the maximum level of the block appears at the output of the selection circuit 42. On the other hand, the minimum level of the current block occurs at the output of the selection circuit 43.

(2) When the detection for the block is completed, the latches 44 and 45 are initialized again.

The dynamic range DR of each block is obtained from the output of the subtraction circuit 46 by subtracting the maximum rail MAX from the selection circuit 42 and the minimum level MIN from the selection circuit 43. These dynamic range DR and minimum level MIN are launched into launches 47 and 48, respectively, by a latch pulse from control block 49. latch 47
The dynamic range DR of each block is output to the output terminal 51 of
is obtained, and the minimum value M I N of each block is obtained at the output terminal 52 of the launch 48.

e. Quantization circuit An example of the quantization circuit 5 will be explained with reference to FIG. In FIG. 8, the bit length Nb (2 bits) determined by the focus length determining circuit 6 from the dynamic range DR is supplied to the input terminal 56, and the minimum value M I N is removed to the input terminal 57. Later pixel data PD
I is supplied. A ROM 55 stores a data conversion table for quantizing these data Nb and PDI.
address input. A 3-bit encoded code DT is obtained from the ROM 55. This encoded code DT is
<Nb=o), all of the 3 bits are invalid; in the case of (Nb"1), the least significant bit of the 3 bits is valid; In the case, 3
The least significant bit and its upper two bits are valid; in the case (Nb=3), all three bits are valid.

In this embodiment, the value of the representative level used as the restoration value is a fixed value depending on the bit length Nb. The dynamic range DR is (0≦DR≦10) and (Nb=O
), there is no need to transmit the encoded code, and the central value (5) of the dynamic range DR is taken as the representative level LO, as shown in the ninth doujinshi. The maximum strain is (EO
=5).

(11≦DR≦25), and in the case of (Nb=1), R
The OM55 performs 1-bit quantization with a maximum distortion of (E 1 =6). As shown in FIG. 9B, if the pixel data PDT is within the range (0 to 12), O'' is output as the encoded code DT, and the pixel data PDT
is within the range (13 to 25), the encoded code DT
"1" is output as . In this case, the representative level used as the restoration value on the receiving side is (L O = 6) (L
1=19).

(26≦DR≦99), and in the case of (Nb=2), R
OM55 performs 2-bit quantization of (E2=12). As shown in FIG. 9C, the pixel data PDI is (0 to 24) and (25 to 49).

(OO), (01), (10), (
11) is output as the encoded code DT. In this case, the representative level used as the restoration value on the receiving side is (L
O=12) (L1=35) (L2=62) (L3=
87).

If (100≦DR≦255) and (Nb=3), the ROM 55 performs 2-bit quantization of (E3=16). As shown in Figure 9, pixel data PD
I is (0-32), (33-65), (66-98),
(99-131), (132-164), (165-1
(000), (001) depending on whether it is in the range of 97L (198-230) or (231-263).
, (010), (011), (100), (Lot
), (110), and (111) are output as encoded codes DT. In this case, the representative levels used as restoration values on the receiving side are (LO=16) (L1=49) (L2
=82)(L3=115)(L4=148)(L5=1
81) (L6=214) (L7=247).

In this way, when the representative level used as a restoration value on the receiving side is a predetermined value corresponding to the dynamic range, it is not necessarily necessary to send the dynamic range DR (8 bits) as an additional code, and the minimum value ( 8 bits) and bit length Nb.

On the other hand, unlike this embodiment, each dynamic range D
Quantization may be performed in a manner that adapts to R and determines the division range each time. In this case, the dynamic range DR is supplied to the ROM as an address signal instead of the pin length Nb. It is assumed that the maximum distortion changes depending on the visual characteristics of the data conversion table for quantization stored in the ROM. In this method, the additional codes are any two of the minimum value (8 bits), dynamic range DR (8 bits), and maximum value (8 bits). Although the latter method lowers the data compression rate, the restoration distortion is significantly improved compared to the above-mentioned methods.

On the receiving side, a representative level determined from the bit length Nb and the encoding code DT is obtained as a restored value, and the minimum value MIN is added to this restored value to obtain pixel data.

! , Modified Example For example, when (Nb=2), as shown in FIG. 10, quantization may be performed so that the representative level LO is equal to the minimum value MIN and the representative level L3 is equal to MAX. good.

Also, one block of data is stored in frame memory.

A combination of a line delay circuit and a sample delay circuit may be used to take out the signals at the same time.

〔Effect of the invention〕

According to this invention, the amount of data to be transmitted can be sufficiently reduced compared to the original data, and the transmission band can be narrowed. Further, the present invention has the advantage that in a stationary portion where the width of change in brightness level is small, the original pixel data can be almost completely restored from the received data, and there is almost no deterioration in image quality. Furthermore, according to the present invention, since the dynamic range is determined for each block, the response is good at transient parts such as edges where the range of change is large.

In this invention, the compression rate can be increased because variable-length encoding is used that does not require a delimiter code.

In particular, in this invention, variable length encoding is performed using nonlinear characteristics that match visual characteristics, so even if the number of bits of the encoding code for each pixel is reduced, deterioration in the quality of the received image due to block distortion etc. can be prevented. be able to.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of the present invention, FIG. 2 is a block diagram showing the configuration of the receiving side, FIG. 3 is a route diagram used to explain blocks that are units of encoding processing, and FIG. 4 , Fig. 5 and Fig. 6 are an example of the configuration of a blocking circuit, a route map and a timing chart for explaining the same, Fig. 7 is a block diagram of an example of a Guinamiso clean range detection circuit, and Fig. 8 is a quantization circuit. A block diagram of an example of the circuit, FIGS. 9 and 10 are route diagrams used to explain one example of quantization and another example, respectively.
FIG. 11 is a route map used to explain the previously proposed high-efficiency encoding. Explanation of main symbols in the drawings 1: Digital television signal input terminal, 2-block conversion circuit, 3: Dynamino clean range detection circuit,
5: Quantization circuit, 6: Binot length determination circuit, 7: Framing circuit. Agent: Tadashi Sugiura, Patent Attorney Chishu Kiwata, 10th grade A゛Figure 1 θ A warm 4f1-l! Figure 2 Figure 3 Figure 10 Figure 9 Figure C'In9 Figure D

Claims (1)

[Claims] The maximum value of a plurality of pixel data contained in a block consisting of an area belonging to the same field or a plurality of consecutive fields of a digital image signal, the minimum value of the plurality of pixel data, and the dynamic range of each block. means for removing the minimum value from the plurality of pixel data for each block, determining a bit length corresponding to the dynamic range and non-linearity for each block, and determining the pixel from which the minimum value has been removed. High efficiency characterized by comprising means for receiving data and performing quantization with the determined bit length, and means for transmitting information representing the dynamic range of each block and the output of the quantization means. Encoding device.