JP3385696B2 - Digital image signal quantizer - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a quantizer used for ADRC encoding a digital video signal, for example.

[0002]

2. Description of the Related Art The applicant of the present application is disclosed in JP-A-61-14498.
A high-efficiency coding apparatus as described in Japanese Patent Publication No. 9 which obtains a dynamic range defined by maximum and minimum values of a plurality of pixels included in a two-dimensional block and performs coding adapted to this dynamic range. Is proposed. Further, as described in Japanese Patent Laid-Open No. 62-92620, there is provided a high-efficiency coding device for coding a three-dimensional block formed from pixels in regions included in a plurality of frames, which is adapted to a dynamic range. Proposed. Further, as described in Japanese Patent Laid-Open No. 62-128621, a variable length coding method is proposed in which the number of bits is converted according to the dynamic range such that the maximum distortion generated when quantizing is constant. Has been done.

In the previously proposed encoding method adapted to the dynamic range (referred to as ADRC), the dynamic range DR (difference between maximum value MAX and minimum value MIN)
Is calculated for each two-dimensional block of the digital video signal composed of, for example, (8 lines × 8 pixels = 64 pixels). Further, the minimum level (minimum value) in the block is removed from the input pixel data. The pixel data after this minimum value removal is quantized.

In this quantization, the detected dynamic range DR is divided into 2 ¹⁶ level ranges corresponding to the number of bits smaller than the original number of quantization bits (for example, 8 bits), for example, 4 bits. This is a process of detecting the level range to which the pixel data belongs and generating a code signal indicating this level range.

When the number of quantization bits is 4, the dynamic range DR of the block is divided into 16 level ranges. Pixel data included in the minimum level range is encoded as (0000), pixel data included in the level range above it is encoded as (0001), and a 4-bit code corresponding to each level range will be described below. The pixel data included in the maximum level range is (11
11) is encoded. Therefore, for example, 8-bit data of each pixel is compressed into 4 bits and transmitted. On the receiving side, the received code signal is restored to the representative level. This representative level is, for example, the central level of each of the 16 level ranges.

It is possible to realize a digital VTR in which a rotary head records / reproduces a digital video signal compressed by the ADRC described above. One of the features of the digital VTR is that there is no image quality deterioration due to multidubbing.

[0007]

However, in actuality, various kinds of processing are performed at the time of encoding or decoding, so that the image quality is deteriorated. For example, an analog video signal is converted into a digital video signal by an A / D converter by a first digital VTR, and after being subjected to a recording process, it is recorded on a tape by a magnetic head, reproduced by a first VTR, and D / A. The converter returns the analog video signal. The reproduced analog video signal is processed by an analog video device that processes the analog video signal and then supplied to the second digital VTR. The second digital VTR performs A / D conversion and recording processing to record the processed signal on the tape, and reproduction processing and D / A conversion processing obtain a reproduced video signal.

As described above, in the dubbing system in which the process of once converting into an analog video signal intervenes, the two digital VTRs are asynchronous and the phases of the sampling clocks for A / D conversion of the respective VTRs do not match.
Due to this phase shift, the video data contained in the block in the second digital VTR corresponding to the block in the first digital VTR is different. This means that the block boundaries are substantially different between the first and second digital VTRs.

A similar problem occurs when the process of converting into an analog video signal is not performed, that is, when dubbing with a digital signal is performed. In an editing system including first and second reproduction digital VTRs, recording digital VTRs, and an editing device, a first reproduction digital VTR is provided.
A part of the image reproduced in step S1 is extracted, and an image in which this is inserted into the image reproduced in the second reproduction digital VTR is created by an editing device, and the edited image is recorded in the recording digital VTR.
Assume a process of recording in TR.

In such a system, when an image is cut out or combined, processing is not always performed at a position that completely coincides with a block boundary. Even if the ADRC block has the same size, the original block is processed. A block shift may occur in which encoding is performed in a block at a slightly displaced position.

FIG. 3 is a diagram for schematically explaining the problems that occur when the above-mentioned block boundaries are shifted. In FIG. 3A, the solid line is the block BL1,
BL2, BL3, ..., Each of which includes a plurality of pixels such as 8 × 8. One pixel X in the block BL1 is ADRC encoded and recorded on the tape. The quantization characteristic Q1 of this encoding is shown in FIG. 3B. In the case of ADRC having a fixed length of n bits, the value of the pixel X is quantized with the quantization width Δ1 obtained by ^dividing the dynamic range of the block BL1 by 1/2 ⁿ .

The data recorded in this way is reproduced,
Be decrypted. This decoded value is L1. Image processing is performed on the decoded data, and the processed data is ADRC.
When coded, a block shift occurs, and the block containing the pixel X is the block B indicated by the broken line in FIG. 3A.
It is assumed to have changed to L5 '. Since this BL5 'is displaced from the block BL1, the dynamic range changes. Therefore, in the second quantization characteristic Q2, the quantization width becomes Δ2 different from Δ1, and the quantization boundary level shifts. The ADRC encoded data is recorded / reproduced as described above, and the reproduced signal is decoded. The level of the decoded output is L2.

When the recording / reproducing process is performed again, even if the quantization characteristic Q3 of the third ADRC is equal to Q1, the decoded output level L3 is the original value X.
It deviates greatly from. As described above, when the block shift exists, the image quality may be deteriorated by repeating the dubbing.

Therefore, an object of the present invention is to provide a quantizer for digital image signals, which has an advantage that deterioration of image quality due to block shift caused by dubbing or the like can be reduced.

[0015]

The present invention provides a block
Is a quantizer that quantizes the digital image signal in the block based on the quantization step width specified in accordance with the dynamic range that is the difference between the maximum value and the minimum value of a plurality of pixels. And possible dynamics
The entire range of the range is divided into multiple ranges, and each range is paired.
Dynamic range candidates and dynamic range
Common quantization step size for each candidate
The minimum value that can be taken for each range candidate is the quantization step.
One or more minimum value candidates with different widths are specified in advance
The input digital image signal block is
Detection means for detecting the dynamic range and minimum value, and detection
Dynamic range that includes the specified dynamic range
A method for determining a dynamic range candidate from the range
And is defined by the determined dynamic range candidates
Among the one or more possible minimum values found,
Means for determining a low minimum most difference small value, is determined
The determined dynamics is determined by subtracting the minimum value
And a quantizing means adapted to quantize by adapting to a range,
Determined dynamic range and minimum and quantizer
It is a quantizer adapted to transmit the output of the stage .

[0016]

[Operation] Dynamic range DR of input digital signal
And the dynamic range DR closest to the minimum value MIN
A combination of 'and the minimum value MIN' is selected, and the quantization is performed accordingly. Since the boundary levels of quantization match among a plurality of combinations, it is possible to prevent the value of the quantized output from varying even when a block shift occurs.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. In FIG. 1, a digital information signal, for example, a digital video signal is supplied to an input terminal designated by 1. In this digital video signal, one sample is digitized into, for example, 8 bits. The input digital signal is converted by the blocking circuit 2 into an array of data from the raster scan order to the block order. One block is a (M pixel × N line) two-dimensional area as a result of subdividing the screen of one frame or one field.

The output signal of the blocking circuit 2 is the detection circuit 3
And the delay circuit 4. The detection circuit 3 detects the maximum value MAX and the minimum value MIN of each block, detects the dynamic range DR that is the difference between MAX and MIN, and outputs the minimum value MIN and the dynamic range DR. The delay circuit 4 delays the data for the time when the detection circuit 3 performs the above detection.

Dynamic range DR and minimum value MI
N is supplied to the determination circuit 5, and the dynamic range DR is supplied to the determination circuit 6. The determination circuit 5 will be described later.
To identify the DR 'corresponding to the dynamic range DR
, And M specified for each DR '
Of the value closest to the detected minimum value MIN in IN '
A discrimination signal SJ 'for identifying MIN' is generated.
The determination signal SJ ′ from the determination circuit 5 is the MIN ′ generation circuit 7
And MIN ′ is generated from this circuit 7. A determination signal SJ is generated from the determination circuit 6, and this signal SJ is DR.
It is supplied to the generation circuit 8 and is also taken out to the output terminal 11.

MIN 'is taken out to the output terminal 12 and supplied to the subtraction circuit 9 to subtract MIN' from the data from the delay circuit 4. The output signal of the subtraction circuit 9 is supplied to the quantization circuit 10. In the quantization circuit 10, D
The output signal of the subtraction circuit 9 is quantized with the quantization step width Δ adapted to R ′. If the quantization bit number is 2 bits, then (Δ = DR ′ × 1/2 ² ). The code signal DT from the quantization circuit 10 is taken out at the output terminal 13.

Although not shown, the coded outputs taken out to the output terminals 11, 12 and 13 are supplied to a framing circuit. The framing circuit adds a sync signal to the encoded output and encodes an error correction code. The transmission data extracted from the framing circuit is recorded on a magnetic tape by, for example, a rotary head.

In the present invention, the boundary level defining the value of the code signal DT from the quantization circuit 10 is restricted so that it can have only a specific value. That is, when the quantization bit number is 2 bits, the combinations are limited to the following.

[0023]

[Table 1]

When the judging circuit 6 receives the dynamic range DR detected by the detecting circuit 3, the judging signal SJ is formed in correspondence with its magnitude. Further, since the minimum value corresponding to each of the dynamic range DR is defined, the determination circuit 5 receives DR and determines DR ′ ,
Indicate the determined DR ', and also to the determined DR'
The minimum value MIN detected among the possible values and the most
Code signal SJ for designating a value MIN 'with a small difference
'Is generated. The MIN 'generating circuit 7 has the above-mentioned Table 1
Is stored, and a code that identifies DR 'and MIN' is stored.
MIN 'is issued by supplying the signal SJ'
Can be born. The judgment circuits 5 and 6 can be composed of a logic circuit, a ROM and the like. Further, the two decision circuits 5 and 6 in FIG. 1 can be combined into one decision circuit.

The reason why the discrimination signal SJ is transmitted is to instruct the decoding device of the value of DR '. 8-bit D
Since the 3-bit discrimination signal SJ is recorded instead of R ′, the recording data amount can be reduced. As for MIN ',
The code signal SJ ′ may be transmitted instead of the 8-bit value, and the amount of recording data can be reduced by combining with the determination signal SJ.

FIG. 2 shows some examples of the quantization characteristics shown in Table 1 above. From the left side of FIG.
= 256, MIN '= 0, Δ = 64), (DR' = 12
8, MIN '= 0, Δ = 32), (DR' = 128, M
IN '= 32, Δ = 32), (DR' = 64, MIN '
= 0, Δ = 16), (DR ′ = 64, MIN ′ = 48,
Δ = 16), (DR ′ = 32, MIN ′ = 0, Δ =
8), (DR '= 32, MIN' = 40, Δ = 8),
The quantization characteristics of (DR ′ = 16, MIN ′ = 0, Δ = 4) are shown. As can be seen from FIG. 2, the quantization boundary levels among the quantization characteristics do not change even if the dynamic range DR differs between blocks.

Although not shown, the receiving (or reproducing) side has a dynamic range DR 'from the discrimination signal SJ.
Alternatively, the quantization step width Δ is generated, the value of the code signal DT is multiplied by Δ, and MIN ′ is added to the multiplication output to perform decoding.

Although the present invention is applied to the fixed length ADRC in the above description, the present invention is applicable to DCT and AD.
It can also be applied to a hybrid code combining RC. Further, the present invention is not limited to ADRC and can be applied to the case of adaptively quantizing a digital signal.

[0029]

According to the present invention, even if the dynamic range is different, the quantization boundary level can be adjusted. Therefore, it is possible to prevent the restoration level from deviating from the original level when the dubbing accompanied by the block shift is performed plural times. Further, in the present invention, it is possible to transmit the code corresponding to these values by utilizing the fact that the dynamic range and the absolute value of the minimum value are defined, and it is possible to reduce the transmission data amount. .

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

FIG. 2 is a schematic diagram showing some specific examples of quantization characteristics in one embodiment of the present invention.

FIG. 3 is a schematic diagram used for explaining a problem that occurs during multidubbing involving block shift.

[Explanation of symbols]

3 detection circuit 5,6 Judgment circuit 10 Quantization circuit

Claims (3)

(57) [Claims] 1. A maximum value and a minimum value of a plurality of pixels in a block
The quantizer step width is specified in accordance with the dynamic range which is the difference between the quantizer step and the quantizer step to quantize the digital image signal in the block based on the quantizer step width . Whole range into multiple ranges
Can be divided into dynamic range candidates corresponding to each range
And a common quantization threshold for each of the above dynamic range candidates.
Take for each step width and the above dynamic range candidates
As the minimum value,
Alternatively, a plurality of minimum value candidates are specified in advance, and a dynamic level is specified for each block of the input digital image signal.
And the dynamics that include the detected dynamic range.
Of the above dynamic range from the range of
Specified in the means of determining and the determined dynamic range candidates
Among the one or more possible minimum values found,
A means for determining the minimum value having the smallest difference from the small value and a plurality of pixels obtained by subtracting the determined minimum value are determined as described above.
Quantization that adapts to the specified dynamic range
Means for determining the dynamic range and the minimum value determined
A quantizer adapted to transmit the output of the quantizing means . 2. The quantizer according to claim 1, wherein instead of transmitting the determined dynamic range, an identification code having a number of bits smaller than the number of bits of the dynamic range is transmitted. Quantizer characterized by. 3. The quantizer according to claim 1, wherein instead of transmitting the determined minimum value , a minimum value
A quantizer characterized by transmitting an identification code having a number of bits smaller than the number of bits .