JP3143970B2 - Image coding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding apparatus, and more particularly to an image encoding apparatus for encoding image data with high efficiency by discrete cosine transform.

[0002]

2. Description of the Related Art When image data is transmitted or recorded on a recording medium such as a magnetic tape, various encodings are employed for compressing image information. For example, so-called predictive coding, transform coding, vector quantization, and the like are known.

[0003] In the above-mentioned transform coding, a sample value (hereinafter referred to as "image data") is converted into axes orthogonal to each other by making use of the correlation of an image signal to de-correlate the correlation between the image data, thereby obtaining the data. The so-called basis vectors are orthogonal to each other, the sum of the average signal powers before conversion and the sum of the average powers of the so-called transform coefficients obtained by the orthogonal transformation are equal, and the power concentration on the low-frequency component is performed. An excellent orthogonal transform is employed. The transform coding includes, for example, so-called Hadamard transform, Haar transform, Karnen-Rube (KL) transform, discrete cosine transform (hereinafter DC)
T: Discrete CosineTransform), discrete sine transform (hereinafter DST: Discrete Sine Transform),
Slant transformation and the like are known.

Here, the DCT will be briefly described. The DCT divides an image into image blocks composed of n (n × n) pixels in both the horizontal and vertical directions in a spatial arrangement, and orthogonally transforms image data in the image blocks using a cosine function. The DCT has been widely used for transmission and recording of image data due to the existence of a high-speed operation algorithm and the realization of a one-chip so-called LSI that enables real-time conversion of image data.
In addition, DCT has almost the same characteristics as the KL transform, which is an optimal transform in terms of the degree of power concentration on low-frequency components that directly affect the efficiency, as the coding efficiency. Therefore, by encoding only the components where power concentrates on the transform coefficients obtained by DCT, it is possible to greatly reduce the amount of information as a whole.

More specifically, n × n image data is converted into DC data.
For example, C _ij (i = 0 to n−
Expressed in 1, j = 0~n-1) , the transform coefficient C ₀₀ corresponds to a DC component representing an average luminance value of the image block, its power is usually much larger than the other components. Therefore, when the coarsely quantizing the DC components, from where visually so-called block distortion is transform coding specific noise felt as a large image quality degradation occurs, the conversion factor C ₀₀
Are assigned a large number of bits (for example, 8 bits or more) and quantized, and transform coefficients C _ij (except C ₀₀ ) of other components (hereinafter, referred to as AC components) except for the DC component have, for example, a visual spatial frequency. Utilizing the visual characteristic of being reduced in the high frequency band, the higher the frequency component, the smaller the number of bits allocated to quantize.

[0006] In transmission and recording of image data, after transform coefficients C _ij obtained by DCT of image data are quantized as described above, so-called Huffman coding (Huffman coding) or the like is used for further compression. Run-length coding (Ru
n Length coding) and the like, transmission and recording are performed by adding a synchronization signal and parity to the obtained coded data.

On the other hand, for example, a digital video tape recorder (hereinafter, referred to as a digital VTR) using a magnetic tape as a recording medium is known as a device for recording a video signal as a digital signal. For example,
As a standardized digital VTR, CCIR
4: 2: 2 digital VTR adopting the so-called 4: 2: 2 component coding standard according to the Rec.
(A so-called D-1 format VTR) and a so-called D-2 format VTR that directly converts a composite signal used for broadcasting into a digital signal and records the digital signal.

The 4: 2: 2 digital VTR converts the luminance signal (Y) and the color difference signals (C _R , C _B ) into 13.
Sampling is performed at a sampling frequency of 5 MHz or 6.75 MHz, converted into a digital signal by 8-bit linear quantization, subjected to predetermined signal processing for adding a synchronization signal, error correction, and the like, and recorded on a magnetic tape. It has become.

Further, in order to reduce the amount of data to be recorded, the 4: 2: 2 component signal is converted into a so-called 4: 1: 1 component signal in which the sampling frequency of the color difference signal is 倍 times, It is possible to record. In other words, the rate conversion from 4: 2: 2 to 4: 1 :: 1 is performed by limiting the band of the chrominance data with a pre-filter, and then thinning out the chrominance data so that the data amount becomes 1/2. By recording a 1: 1 component signal (data), the data amount can be reduced.

[0010]

By the way, for example, the above-described DCT coding and the 4: 1 coding on a digital VTR are performed.
It is conceivable that the two techniques using one component signal are applied at the same time and recording is performed with the data amount significantly reduced as compared with the conventional apparatus. However, in addition to the circuit for DCT, A filter, for example, a digital low-pass filter and a data processing circuit for thinning out color difference data are required, and there is a disadvantage that the circuit scale is increased.

The present invention has been made in view of such circumstances, and in addition to coding by discrete cosine transform, data equivalent to rate conversion from, for example, 4: 2: 2 to 4: 1: 1. It is an object of the present invention to provide an image encoding apparatus capable of realizing the effect of reducing the amount without requiring a pre-filter and a data processing circuit for thinning out data.

[0012]

According to the present invention, in order to solve the above-mentioned problems, there are provided block-forming means for dividing luminance data and chrominance data into blocks each having n × n blocks in a spatial arrangement. Discrete cosine transform means for orthogonally transforming the luminance data of each block and the chrominance data of each block from the blocking means using a cosine function to calculate a transform coefficient of luminance data and a transform coefficient of chrominance data; High-frequency component removing means for setting the value of the high-frequency component of the horizontal half of the conversion coefficient of the color difference data from the conversion means to zero, the conversion coefficient of the luminance data from the discrete cosine conversion means and the high-frequency component And a quantizing means for quantizing the transform coefficient of the color difference data from the removing means to form quantized data, and outputting the quantized data.

[0013]

In the image coding apparatus according to the present invention, the luminance data and the chrominance data are divided into blocks each having n × n blocks in a spatial arrangement, and the luminance data of each block and the chrominance data of each block are respectively divided. The transform coefficient of the luminance data and the transform coefficient of the chrominance data are calculated by performing orthogonal transform using the cosine function, and the value of the high-frequency component of １ ／ of the transform coefficient of the chrominance data in the horizontal direction is set to zero. Then, the transform coefficient of the luminance data and the transform coefficient of the chrominance data with the high-frequency component set to zero are quantized to form quantized data, and these quantized data are output.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an image coding apparatus according to the present invention will be described below with reference to the drawings. FIG. 1 shows a circuit configuration of an image encoding apparatus to which the present invention is applied, and FIG. 2 is a circuit configuration of a recording system of a digital video tape recorder (hereinafter referred to as a digital VTR) to which the image encoding apparatus is applied. FIG. 3 shows a circuit configuration of a reproduction system of the digital VTR.

First, the digital VTR will be described. As shown in FIG. 2, the digital VTR performs data compression such as so-called conversion coding on the luminance data Y and the color difference data C _R and C _B supplied as so-called 4: 2: 2 component signals to compress data. After that, a recording system for recording on the magnetic tape 1 via the magnetic head 21 and, as shown in FIG. 3, a reproduction signal reproduced from the magnetic tape 1 by the magnetic head 31 are binarized and then decoded. And a reproducing system that reproduces the luminance data Y and the color difference data C _R and C _B by performing the data processing described above.

The above recording system, as shown in FIG.
An image block G _Yh , in which n × n blocks in the spatial arrangement of the luminance data Y and the color difference data C _R and C _B are one block,
A blocking circuit 12 that divides the image into G _Rh and G _Bh and forms a processing unit including a predetermined number of image blocks at a predetermined ratio;
Blocked luminance data Y and color difference data C _R from the orthogonal transformation by using respectively the cosine function C _B (hereinafter DCT: Discrete referred Cosine Transform) to the image block G _Yh, G _Rh, each transformation G _Bh Coefficient C _ij (i = 0 to
DCT circuit 13 for calculating (n-1, j = 0 to n-1)
The value of the horizontal high-frequency component of each of the conversion coefficients C _ij of the color difference data C _R and C _B from the DCT circuit 13 is set to zero, and each of the conversion coefficients C _{ij of the} color difference data C _R and C _B. And D above
A quantization circuit 14 that quantizes the transform coefficient C _ij of the luminance data Y from the CT circuit 13 for each processing unit to form quantized data of each data, and a quantized data from the quantizer 14 is, for example, a so-called An encoding circuit 15 for encoding with a variable length code to form encoded data VLC _ij (i = 0 to n−1, j = 0 to n−1); and encoded data VLC _ij from the encoding circuit 15 For example, a parity addition circuit 17 for adding a parity for error detection and error correction for each processing unit, and a synchronization signal and the like are added to the encoded data VLC _ij to which the parity from the parity addition circuit 17 is added. A synchronization signal insertion circuit 18 for adding transmission data to each transmission data to form transmission data; and a parallel / serial converter for converting transmission data transmitted from the synchronization signal insertion circuit 18 as parallel data into serial data. Al (hereinafter referred to as P / S) converter 19
And a channel encoder (hereinafter referred to as ENC) 20 that performs a modulation process suitable for recording on the transmission data from the P / S converter 19 to generate a recording signal, and supplies the recording signal to the magnetic head 21. Be composed.

In this recording system, the luminance data Y and the chrominance data C _R and C _B supplied as 4: 2: 2 component signals via the terminal 2 are, for example, 8 × 8 in a spatial arrangement as one block. Image block G _Yh ,
G _Rh and G _Bh , respectively, and the luminance data Y of the image block G _Yh and the color difference data C _R and G _{Rh of} the image blocks G _Rh and G _Bh are obtained.
The C _B each DCT to calculate the transform coefficients C _ij. Then, color difference data C _R, the value of the horizontal direction, for example, half of the high frequency component of the transformation coefficients C _ij of C _B is set to zero, the conversion of the color difference data C _R, transform coefficients C _B C _ij and the luminance data Y The coefficients are quantized and quantized for each processing unit that is a ratio of 4: 2: 2 and includes a predetermined number of image blocks G _Yh , G _Rh , and G _Bh , for example, one unit of data processing or transmission. In addition to forming data, the quantized data is encoded by a variable length code to form encoded data VLC _ij . Further, this recording system uses the encoded data V
After adding a synchronization signal or the like to the LC _ij for each processing unit to form transmission data, the transmission data is subjected to modulation suitable for recording, for example, scrambling or NRZI modulation processing, and is recorded on the magnetic tape 1 by the magnetic head 21. It has become.

Thus, the image coding apparatus according to the present invention,
That is, the main part of the digital VTR configured as described above is composed of the blocking circuit 12 to the quantization circuit 14, and the blocking circuit 12 to the quantization circuit 14
Luminance data Y and color difference data C _R, respectively, 2 for C _B
It has two processing systems, and is specifically configured as follows.

The blocking circuit 12 has a storage capacity of, for example, one field or one frame and stores the luminance data Y, as shown in FIG.
and an image block G _Yh (h = 1 to H, where H = 1 to H, where H is the number of pixels in one frame or one field and one block of pixels) while divided into depends on the number n ^2), the luminance data Y and color difference data C _R, C _B
Is a ratio of 4: 2: 2, for example, and the blocking unit 12b reads out a predetermined number of image blocks G _Yh.
And a memory 12c having a storage capacity of, for example, one field or one frame and storing the color difference data C _R and C _B , respectively, and the color difference data C _R and C from the memory 12c.
_B is divided into image blocks G _Rh and G _Bh (h = 1 to H, where H is に 対 し て with respect to the luminance data Y) each having n × n blocks in the spatial arrangement as one block, and each is divided into predetermined blocks. And a blocking unit 12d for reading out each of the number of image blocks G _Rh and G _Bh .

[0020] Then, the blocking circuit 12, the image data supplied via respective terminals 2, i.e. 4: 2: luminance data Y supplied 2 ratio and color difference data C _R, a C _B respectively memory 12a, 12c, and temporarily stores the luminance data Y stored in the memory 12a.
_Is an image block G _Yh with 8 × 8 blocks as one block, for example.
And the color difference data C stored in the memory 12c.
_R, the image block G _Rh to each example 8 × 8 pieces of one block C _B, as well as divided into G _Bh, 4: 2: 2
And a predetermined number of image blocks G _Yh , G
_Rh and G _Bh are read for each processing unit, and the read luminance data Y and color difference data C _R and C _B are read by the DCT circuit 1.
3.

The DCT circuit 13 includes, for example, a so-called DSP
(Digital Signal Processor) and the like, and luminance data Y supplied from the blocking circuit 12 for each processing unit.
13a to calculate the conversion coefficient C _ij of the luminance data Y by orthogonally transforming the color data by using a cosine function, and the color difference data C _R and C _B supplied from the blocking circuit 12 for each processing unit.
Are respectively orthogonally transformed using a cosine function to obtain color difference data C.
DCT1 for calculating respective transform coefficients C _ij of _R and C _B
3b.

[0022] Then, the DCT circuit 13, the luminance data Y and color difference data C _R supplied to each processing unit from the blocking circuit 12, and DCT to C _B, transform coefficients obtained luminance data Y C _ij and The conversion coefficients C _ij of the color difference data C _R and C _B are supplied to the quantization circuit 14.

As shown in FIG. 1, the quantization circuit 14 has different quantization widths from each other, and
a quantizer Q _Ym (m = 1 to M) for quantizing the transform coefficient C _ij of the luminance data Y from a to form quantized data of different data amounts for the same processing unit,
A zero setting circuit 14a for setting the value of the high-frequency component in the horizontal direction of each of the conversion coefficients C _ij of the color difference data C _R and C _B from the DCT 13b to zero; The quantizer Q _Cm (m = 1 to M) that quantizes the transform coefficients C _ij of the color difference data C _R and C _{B from} the image processing unit 14a and forms quantized data of different data amounts for the same processing unit. ).

The quantization circuit 14 converts the conversion coefficient C _ij of the luminance data Y and the color difference data C
_R, forming each transform coefficient C _ij of C _B, the same processing of the luminance data Y of different data amounts respectively unit quantized data and color difference data C _R, for each processing unit to each quantized data C _B Then, these quantized data are encoded by the encoding circuit 15.
To be supplied.

More specifically, as shown in FIG. 4, for example, as shown in FIG. 4, the zero setting circuit 14a generates a half of the high frequency component in the horizontal direction in the area 50 of the conversion coefficients C _ij of the color difference data C _R and C _B. The value of the transform coefficient C _{ij in} the area 51 is forcibly set to zero, and the transform coefficient C _ij thus obtained is supplied to the quantizer _QCm .

Each of the quantizers Q _Ym and Q _Cm has a different quantization width. For example, the quantizers Q _Y1 and Q _C1 have a predetermined quantization width q, and the quantizers Q _Y2 and Q _Cm have a predetermined quantization width q. _C2 is quantization width 2
q, and the quantizers Q _Y3 and Q _C3 have a quantization width of 4q,
_.. , The quantizer Q _Ym forms quantized data of luminance data Y having different data amounts for the same processing unit, and the quantizer Q _Cm color difference data C _R different data amount from one another for, respectively forming each quantized data C _B, has these quantized data is supplied to the encoding circuit 15.

As shown in FIG. 1 described above, the encoding circuit 15 includes, for example, a so-called Huffman code (Huffman code) for performing variable length coding and a run length code (Run Length c).
consists ode), etc., and which encoded by a variable length coding the quantized data of the luminance data Y from the respective quantizer Q _Ym, coder COD _Ym to form the encoded data VLC _ij of luminance data Y, respectively (M = 1 to M) and a buffer memory BUF _Ym (m = 1 to M) that stores the encoded data VLC _ij from each encoder COD _Ym for each processing unit and has a predetermined storage capacity. , The respective buffer memories B
Encoded data VLC _ij read from UF _Ym
And the selector 15a for detecting one of the buffer memories BUF _Ym and detecting the overflow of each of the buffer memories BUF _Ym.
and a control circuit 15b for controlling a, likewise made from the Huffman coder and run length coder or the like, each quantizer Q _Cm
, And each of the quantized data of the color difference data C _R and C _B is encoded by a variable length code, and the color difference data C _R and C _B are respectively encoded.
_Encoders respectively forming the encoded data VLC _ij of
OD _Cm (m = 1 to M) and each encoded data VLC _ij from each encoder COD _Cm are stored for each processing unit,
Buffer memory BUF _Cm having a predetermined storage capacity (m =
And 1 to M), detects a selector 15c for selecting one of the respective read encoded data VLC _ij from respective buffer memory BUF _Cm, the overflow of the respective buffer memories BUF _Cm respectively, quantization to be described later Control circuit 15d for controlling the selector 15c by a device selection signal
A MUX that time-division multiplexes the conversion coefficient C _ij of the luminance data Y selected by the selector 15a and the conversion coefficients C _{ij of} the color difference data C _R and C _B selected by the selector 15c.
15e.

The encoding circuit 15 encodes the quantized data of the luminance data Y having different data amounts from the respective quantizers Q _Ym by using a Huffman code and a run-length code, and applies the same to the same processing unit. To form the encoded data VLC _ij of the luminance data Y having different data amounts from each other, store these encoded data VLC _ij in the buffer memory BUF _Ym , respectively, and detect overflow of the buffer memory BUF _Ym , A quantizer selection signal for selecting the quantizer Q _Ym that does not cause overflow and has the maximum data amount, that is, the number m _Y of the quantizer Q _Ym is supplied to the selector 15a, and is selected by the selector 15a. The encoded data VLC _ij of the luminance data Y is supplied to the MUX 15e.

The encoding circuit 15 also generates color difference data C _R , C _R of different data amounts from the respective quantizers Q _Cm.
Each of the quantized data of _B is encoded by a Huffman code and a run-length code to form encoded data VLC _ij of color difference data C _R and C _B having different data amounts for the same processing unit, respectively. these encoded data VLC _ij stores respectively in the buffer memory BUF _Cm, to detect an overflow of the buffer memory BUF _Cm, without causing an overflow, and selects a quantizer Q _Cm having the maximum amount of data quantizer select signal for, i.e. the number m _C quantizer Q _Cm is supplied to the selector 15c, color difference data C _R selected by the selector 15c, each coded data VLC _ij of C _B MU
X15e. And MUX
15e is time-division-multiplexes the encoded data VLC _ij and color difference data C _R of the luminance data Y, each coded data VLC _ij of C _B, via the terminal 5 and outputs the parity adding circuit 17 shown in FIG. 2 above It has become. The encoding circuit 15 also assigns the numbers m _Y and m _C of the quantizers Q _Ym and Q _Cm selected by the selectors 15 a and 15 c to the parity shown in FIG. The signal is supplied to the additional circuit 17.

As a result, from the MUX 15e, the luminance data Y and the chrominance data C _R obtained by being quantized with the minimum quantization width so that the data amount of the processing unit is within a predetermined amount and the data amount is maximized. , C _B coded data VLC
_ij is output. At this time, the high-frequency component (1/2 of the whole) in the horizontal direction of each of the quantized data of the chrominance data C _R and C _B is set to zero, as described above, for example, so that the chrominance described in the related art is used. After limiting the bands of the data C _R and C _B with the pre-filter, the data is thinned out so that the data amount of the color difference data C _R and C _B is reduced to 行 う.
An effect of reducing the data amount equivalent to the rate conversion to 1: 1 can be obtained. In other words, it does not require a pre-filter and a data processing circuit for thinning out the data unlike the conventional device, and can reduce the amount of data equivalent to the rate conversion from 4: 2: 2 to 4: 1: 1. It can be carried out.

The parity addition circuit 17 and the synchronization signal insertion circuit 18 shown in FIG. 2 correspond to the selected quantizers Q _Ym and Q _Cm numbers m _Y and m _C from the encoding circuit 15 and the luminance data. Y and each encoded data V of color difference data C _R and C _B
The LC _ij is time-division multiplexed, and parity, a synchronization signal and the like are added to form transmission data. As a result, for example,
One processing unit is a synchronization signal in order from the top, the numbers m _Y and m _{C of} the quantizers Q _Ym and Q _Cm employed in the processing unit, and the codes of a predetermined number of image blocks G _Yh , G _Rh and G _Bh , respectively. The transmission data including the encoded data VLC _ij and the parity is output.

[0032] As described above, the image coding apparatus, an image block G _Yh to n × n pieces one block in each spatial arrangement luminance data Y and color difference data C _R, a C _B, G _Rh, G _Bh And each image block G
Luminance data Y of _Yh , color difference data C of each image block G _Rh
R and the color difference data C _B of each image block G _Bh each orthogonal transform using a cosine function to calculate transform coefficients of the luminance data Y C _ij and color difference data C _R, each transform coefficient C _ij of C _B, the The high-frequency component in the horizontal direction of each of the conversion coefficients C _ij of the color difference data C _R and C _B is set to zero. Then, the conversion coefficient C _ij of the luminance data Y and the color difference data C with the high frequency component set to zero
_R, luminance data Y and color difference data C _R each transform coefficient C _ij by quantization of C _B, by forming each quantized data C _B, the thinning of the prefilter and the data as in the conventional apparatus No data processing circuit is required, for example, 4:
It is possible to reduce the data amount equivalent to the rate conversion from 2: 2 to 4: 1: 1.

Next, the reproduction system of the digital VTR will be described. This reproducing system, as shown in FIG.
A channel decoder (hereinafter simply referred to as DEC) 32 for performing signal processing such as NRZI demodulation on a reproduction signal reproduced from the magnetic tape 1 by the magnetic head 31 to reproduce transmission data, and transmission transmitted as serial data from the DEC 32 A serial / parallel (hereinafter, referred to as S / P) converter 33 for converting data into parallel data, and synchronizing transmission data from the S / P converter 33, as well as luminance data Y and color difference data C _R and C _B. each coded with the synchronization signal detection circuit 34 for reproducing data VLC _ij, time base correction circuit for correcting the variation of the time axis occurring during playback of each of the encoded data VLC _ij (hereinafter the TBC: time Base
Corrector) 35 and error correction of each encoded data VLC _ij from the TBC 35.
An error correction circuit 36 for setting an error flag EF for the encoded data VLC _ij for which error correction has failed;
At the time of recording from the error correction circuit 36, each encoded data VLC _ij which has been
And color difference data C _R, a decoding circuit 37 for reproducing the quantized data C _B, color difference data C from該復Goka circuit 37
_R, C values for each quantized data _B are added to the data is zero as a horizontal high-frequency component, this way is formed by the color difference data C _R, quantized data and the decoding circuit of C _B 37 Performs inverse quantization signal processing on the quantized data of the luminance data Y from the luminance data Y and the color difference data C _R , C
An inverse quantization circuit 38 for reproducing each of the conversion coefficients C _ij of _B , and an inverse quantization circuit for orthogonally transforming each of the conversion coefficients C _ij from the inverse quantization circuit 38 to reproduce the luminance data Y and the color difference data C _R and C _B. An orthogonal transform circuit (hereinafter referred to as an IDCT circuit) 39;
From the T circuit 39, the image blocks G _Yh , G _Rh , G _Bh
Inverse blocking circuit 40 to form the luminance data Y, color difference data C _R, C _B luminance data of one frame or one field from Y and color difference data C _R, C _B supplied per
And the luminance data Y from the deblocking circuit 40 based on the error flag EF from the error correction circuit 36.
And an error correction circuit 41 for performing error correction on the color difference data C _R and C _B respectively.

Next, the operation of the reproducing system configured as described above will be described. The DEC 32 binarizes a reproduction signal reproduced by the magnetic head 31 from the magnetic tape 1, performs NRZI demodulation, performs a descrambling process, reproduces transmission data, and converts the transmission data to S / S.
The signal is supplied to the synchronization signal detection circuit 34 via the P converter 33.

The synchronizing signal detecting circuit 34 includes the S / P converter 3
3, a synchronization signal is detected from the transmission data converted to the parallel data to synchronize the data, and the encoded data VLC _ij of the luminance data Y and the color difference data C _R and C _B are reproduced. _ij is supplied to the TBC 35.

The TBC 35 corrects the time axis of each of the encoded data VLC _ij , absorbs the fluctuation of the time axis generated at the time of reproduction, and obtains the time axis corrected coded data VLC _ij.
Is supplied to the error correction circuit 36.

The error correction circuit 36 outputs each encoded data V
The error correction of the LC _ij is performed using the parity added at the time of recording, and the error flag EF is set for the encoded data VLC _ij having an error exceeding the error correction capability, and the error-corrected luminance data is set. The decoding circuit 37 converts the encoded data VLC _ij of Y and the color difference data C _R and C _B into
To supply.

The decoding circuit 37 decodes each of the encoded data VLC _ij encoded by the Huffman code and the run length code at the time of recording, and performs each quantization of the luminance data Y and the chrominance data C _R and C _B. The data is reproduced, and the quantized data is supplied to the inverse quantization circuit 38.

The inverse quantization circuit 38 calculates each processing unit used at the time of recording based on the number m _Y of the quantizer Q _Ym of each processing unit reproduced together with the encoded data VLC _ij of the luminance data Y. It recognizes the quantizer Q _Ym of reproducing transform coefficients C _ij of luminance data Y by the inverse quantization respectively quantized data of the luminance data Y of each processing unit in the quantization width corresponding to the quantizer Q _Ym The conversion coefficient C _ij is supplied to the IDCT circuit 39. Further, the inverse quantization circuit 38 is the value at the time of recording is zero, that do not play color difference data C _R, as a high-frequency component value is zero data of each quantized data C _B, n
× n pieces of color difference data C _R, after forming the respective quantized data C _B, color difference data C _R, each encoded data VLC of C _B
The number m of the quantizer Q _Cm of each processing unit reproduced together with _ij
Based and _C, recognizes the quantizer Q _Cm of each processing unit that is used during recording, color difference data C _R of each processing unit in the quantization width corresponding to the quantizer Q _Cm, the C _B Each quantized data is inversely quantized to reproduce each transform coefficient C _ij of the color difference data C _R and C _B , and these transform coefficients C _ij are identified by IDCT.
Supply to circuit 39.

The IDCT circuit 39 orthogonally transforms each of the conversion coefficients C _ij of the luminance data Y and the chrominance data C _R and C _B using a transposed matrix corresponding to the conversion matrix used at the time of recording to perform luminance data conversion. Y and for each image block G _Yh, color difference data C _R, each image block G _Rh and C _B, reproduced for each G _Bh, these luminance data Y and color difference data C _R, a C _B in the reverse blocking circuit 40 Supply.

The deblocking circuit 40 generates the image block G
_The luminance data Y, supplied for each of _Yh , G _Rh and _GBh ,
Color difference data C _R, the luminance data Y for one frame or one field from C _B and color difference data C _R, and supplies to form a C _B to the error correction circuit 41.

The error correction circuit 41 performs an interpolation process for each of the luminance data Y and the chrominance data C _R and C _B using, for example, data having no error in the vicinity of the data for which the error correction circuit 36 has failed to correct the error. By doing so, the luminance data Y and the color difference data C _R , C
The error correction of _B is performed, and the luminance data Y and the chrominance data C _R and C _{B with} the error corrected, that is, the luminance data Y and the chrominance data C in the 4: 2: 2 component signal.
_R, and outputs the C _B.

As described above, the data amount of a processing unit composed of a predetermined number of image blocks G _Yh , G _Rh , and G _Bh at a ratio of 4: 2: 2 is set to a fixed value or less (fixed length). When performing quantization and recording so that quantization distortion is minimized within the data amount, the value of the high-frequency component of each conversion coefficient C _ij of the color difference data C _R and C _B is set to zero,
Color difference data C that does not significantly affect image quality during recording
_R, since the efficient coding high-frequency components of C _B as zero, it is possible to obtain a reproduced signal of good quality. Further, since the processing unit has a fixed length, editing, variable speed reproduction, and the like can be easily performed. Also, color difference data C _R not play to the value set to zero during recording, the high frequency components of the quantized data C _B, upon dequantization, that value is added as data zero, the color difference Data C _R , C _B
Can be reproduced, and unlike the conventional apparatus, it is not necessary to generate the thinned color difference data by interpolation processing or the like, and the circuit scale of the reproduction system can be reduced.

[0044]

As is clear from the above description, according to the present invention, luminance data and chrominance data are converted into nx.times.
n is divided into blocks each having one block, and the luminance data of each block and the chrominance data of each block are orthogonally transformed using a cosine function to calculate a conversion coefficient of luminance data and a conversion coefficient of chrominance data, The value of the high-frequency component of 変 換 of the conversion coefficient of the color difference data in the horizontal direction is set to zero. Then, the transform coefficient of the luminance data and the transform coefficient of the chrominance data with the high-frequency component set to zero are quantized to form quantized data, and these quantized data are output. No data processing circuit is needed for thinning out the filter and data, and 4: 2: 2 to 4:
It is possible to reduce the amount of data equivalent to the rate conversion to 1: 1 and to obtain high coding efficiency.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a circuit configuration of an image encoding device to which the present invention has been applied.

FIG. 2 is a block diagram showing a circuit configuration of a recording system of a digital video tape recorder to which the image encoding device is applied.

FIG. 3 is a block diagram showing a circuit configuration of a reproduction system of a digital video tape recorder to which the image encoding device is applied.

FIG. 4 is a diagram illustrating an area of a high-frequency component in a horizontal direction of a conversion coefficient of color difference data.

[Explanation of symbols]

12 Blocking circuit 13 DCT circuit 14a Zero setting circuit Q _Ym , Q _Cm Quantizer

Claims (1)

(57) [Claims] 1. Blocking means for dividing luminance data and chrominance data into blocks each having n × n blocks in a spatial arrangement, and luminance data of each block and chrominance data of each block from the blocking means. Are respectively orthogonally transformed by using a cosine function to calculate a conversion coefficient of luminance data and a conversion coefficient of chrominance data, and １ ／ of a conversion coefficient of chrominance data from the discrete cosine conversion means in a horizontal direction. A high-frequency component removing means for setting the value of the high-frequency component to zero, a transform coefficient of the luminance data from the discrete cosine transform means and a transform coefficient of the color difference data from the high-frequency component removing means, and quantized data. , And a quantizing means for outputting the quantized data.