JPH07111633A - Image compression recording/reproducing device - Google Patents

Abstract

PURPOSE:To improve compression efficiency by maintaining interchangeability by a long time mode and a standard mode by performing a variable length encoding in basic data and a fixed length encoding in additional data according to the nature of data. CONSTITUTION:The data inputted from an input terminal 1 is divided into an 8X8 DCT block by a blocking circuit 2. The data is converted into 64 DCT coefficients every each block in a DCT circuit 3 and is outputted in order of a zigzag scan. A linear quantization is performed for the DCT coefficients by the quantization value multiplied by different quantization steps and scaling factors for each coefficient in a quantizer 4 and the coefficients are outputted. In a separation circuit 5, a bit assignment is performed according to a quantization characteristic. Outputted upper bit data is recorded in a magnetic tape 14 by heads 12 and 13 after a signal processing is performed for the data in a circuit 7 after a variable length encoding is performed for the data in a VCL circuit 6. A signal processing is performed for the lower bit data from the circuit 5 in a circuit 9 after a fixed length encoding is performed for the data in an FLC circuit 8 and the data is recorded by heads 14 and 15.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image compression recording / reproducing apparatus for compressing and encoding digitized image data for recording / reproducing.

[0002]

2. Description of the Related Art In recent years, due to the progress of image compression coding technology and the development of high-density magnetic recording, a compression type VTR for compressing and coding image data and digitally recording it has been proposed and is nearing practical use. In addition, in order to meet various demands such as long-term recording and high image quality recording, a standard mode that achieves high image quality without reducing the amount of information with data at a low compression ratio and an information amount with data at a high compression ratio are provided. A method has been proposed in which recording is performed by switching the long-time mode, which is intended for long-term recording with a significant reduction.

Further, from the viewpoint of downsizing and cost reduction of equipment, VT compatible with only one of these methods.
Since R is also considered, compatibility between them becomes a problem.

Therefore, a VTR that records high-compression information as a long-time mode and additionally records information that can improve image quality by adding to the high-compression information in the standard mode is recorded. Proposed (JP-A-3
-121684).

In this VTR, an orthogonal transform is used as a compression coding method. Then, in the long-time mode, bits are mainly allocated to important DC and low-frequency AC components of the coefficient as basic data, and in the standard mode, more high-frequency AC components are allocated bits to add data other than the basic data. Data is recorded as data.With these configurations, it is possible to configure a system that is compatible with both long-time mode and standard mode, and it is possible to play back tapes recorded in standard mode even on long-time mode dedicated machines. ing.

[0006]

However, in the above-mentioned conventional configuration, only fixed-length coding is possible,
Variable-length coding with higher compression efficiency cannot be directly supported. That is, in a compression method in which a variable length code is combined with an orthogonal transformer represented by JPEG or the like, the code amount after encoding varies depending on the characteristics of image data, but in a VTR, one screen is reduced to a predetermined code amount or less. This is because it needs to be encoded and needs to be adaptively quantized according to the characteristics of the image data.

Although it is conceivable that both standard data and high image quality additional data are variable length coded and recorded, two systems of code amount control, variable length coding circuit, etc. are required, which causes an increase in circuit scale. Is coming.

The present invention solves the above-mentioned conventional problems, and it is possible to improve the compression efficiency in both the long-time mode and the standard mode, and at the same time, the image compression recording / reproduction capable of maintaining the compatibility. The purpose is to provide a device.

[0009]

In order to achieve this object, an image compression recording / reproducing apparatus according to the present invention comprises an orthogonal transformer for orthogonally transforming input image data block by block, and an output data of the orthogonal transformer. A quantizer, and a separating circuit for separating the quantizer output into upper bits and lower bits,
A variable-length encoder for variable-length encoding the upper bit output of the separation circuit, a first recording head for recording the variable-length encoded data on a recording medium, and a fixed-length code for the lower bit output of the separation circuit. A fixed-length encoder for encoding, a second recording head for recording the fixed-length encoded data on the recording medium, a first reproducing head for reproducing variable-length encoded data from the recording medium, A variable length decoder for variable length decoding the variable length encoded data, a second reproducing head for reproducing fixed length encoded data from the recording medium, and a fixed length decoding for the reproduced fixed length encoded data. A fixed length decoder, a combining circuit for combining data from the variable length decoder output and the fixed length decoder output, an inverse quantizer for inverse quantizing the combining circuit output, and the inverse quantizer Inverse orthogonal transform of output and image An inverse orthogonal transformer that outputs data, and at the time of recording, only the variable length encoded data is recorded in the first recording mode, and the variable length encoded data and the fixed length are recorded in the second recording mode. When both the encoded data are recorded and reproduced, the synthesizing circuit, when the recording medium is recorded in the first recording mode, sets the upper bit to the variable length decoder output and the lower bit to 0, and when the recording medium is recorded in the second recording mode, the upper bits are output as the variable length decoder output, and the lower bits are output as the fixed length decoder output. There is.

[0010]

With this structure, the basic data is compressed by the variable-length coding and the additional data is compressed by the fixed-length coding, so that the efficient compression can be performed in any mode. While ensuring the compatibility of, improve the quality of the reproduced image.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram of an image compression recording / reproducing apparatus according to an embodiment of the present invention, FIG. 2 is a diagram showing a zigzag scan order of DCT, FIG. 3 is a diagram showing a quantization step of each coefficient of DCT, and FIG. FIG. 5 is a diagram showing allocation of lower bits, FIG. 5 is a diagram showing an operation of the separation circuit, FIG. 6 is a diagram showing a tape pattern, and FIG. 7 is a diagram showing an operation of the synthesizing circuit.

In FIG. 1, 1 is an input terminal for inputting image data, 2 is a blocking circuit for dividing the image data into 8 × 8 blocks, and 3 is an example of orthogonal transformation for each 8 × 8 block. DCT for cosine transform (DCT)
A circuit, 4 is a quantizer for quantizing DCT coefficients, 5 is a separation circuit for separating the quantizer output into upper bits and lower bits, and 6 is a VL for variable length encoding (VLC) the upper bits.
C circuit, 7 is a first recording processing circuit that converts VLC data into a signal suitable for recording, 8 is a FLC circuit that performs fixed length encoding (FLC) of lower bits, and 9 is a signal suitable for recording FLC data A second recording processing circuit for conversion, 10 is a mode input terminal for inputting a recording mode, 11 is a mode change switch for switching the output of the second recording processing circuit according to the recording mode, and 12 and 13 are basic data recording heads. , 14 and 15 are heads for recording additional data, 16
Is a magnetic tape.

Reference numerals 17 and 18 are basic data reproducing heads, 19 and 20 are additional data reproducing heads, 21 is a mode detection circuit for detecting a recorded mode, and 22 is a first reproduction for converting a head reproduction signal into data. Processing circuit,
Reference numeral 23 is a VLD circuit for decoding (VLD) VLC data and outputting upper bits, 24 is a second reproduction processing circuit for converting a head reproduction signal into data, and 25 is decoding (FLD) FLC data for lower bits. FLD circuit for output, 2
6 is a synthesizing circuit for synthesizing the upper bit and the lower bit, 27
Is an inverse quantization circuit for inversely quantizing the combined data, 28
Is the ID that inversely transforms the inversely quantized data by discrete cosine
A CT circuit, 29 is an inverse blocking circuit for rearranging blocked data into image data, and 30 is an output terminal for outputting image data.

The operation of the image compression recording / reproducing apparatus configured as described above will be described below. First, the data input from the input terminal 1 is divided into 8 × 8 DCT blocks by the blocking circuit 2. Data is DCT
The circuit 3 converts each block into 64 DCT coefficients and outputs the DCT coefficients in the zigzag scan order shown in FIG. DCT
The coefficient is linearly quantized by the quantizer 4 and output by a quantized value obtained by multiplying a quantization step different for each coefficient shown in FIG. 3 and a scaling factor. Here, as the scaling factor, a value is selected so that the code amount after variable-length coding falls within a predetermined code amount or less.

In the separation circuit 5, first, bit allocation is performed according to the quantization characteristic. When the scaling factor is small, for example, if the scaling factor is 1/4, the quantized value in the low frequency band will be 1/4, so only the upper bits will be DC.
The T coefficient is represented, and no quantization distortion occurs. Therefore, FIG.
It is possible to efficiently improve the image quality by assigning the lower bits to the high-frequency DCT coefficient having a large quantization step such as a. Further, when the scaling factor is large, the distortion due to quantization affects all the coefficients, so that the lower bits are allocated over the entire range of the DCT coefficients as shown in FIG. 4b.

Next, as shown in FIG. 5, data excluding the lower 2 bits of the quantized data is rounded and output as the upper bits, and the number of bits of the lower 2 bits according to the bit allocation is converted into the lower bits. Output as. The output high-order bit data is variable-length coded by the VLC circuit 6, and then a signal suitable for magnetic recording such as shuffling, addition of error correction code, and recording coding is performed by the first recording processing circuit 7. Processed and output. Then, the data is recorded on the magnetic tape 14 by the basic data recording heads 12 and 13.

In the FLC circuit 8, the lower bit data output from the separation circuit 5 is fixed-length coded, and then V
Similar to the LC data, the second recording processing circuit 9 performs signal processing suitable for magnetic recording and outputs. Then, the mode changeover switch 11 is changed over according to the mode signal from the mode input terminal 10, and recording is performed by the additional data recording heads 14 and 15 only in the standard mode.

The recorded tape pattern is shown in FIG.
In the long time mode, only the basic data is recorded as one track per field, and in the standard mode, the basic data and the additional data are alternately recorded in a total of two tracks per field. Therefore, in the long time mode, the recording time which is twice as long as that in the standard mode can be secured.

The signals reproduced by the basic data reproducing heads 17, 18 are subjected to reproduction processing such as reproduction decoding, error correction, deshuffling, etc. in the first reproduction processing circuit 22 and output as VLC data. . The VLD circuit 23 performs variable length decoding processing on the VLC data and outputs it as upper bit data. In the case of standard mode recording,
The signals reproduced by the additional data reproducing heads 19 and 20 are reproduced by the second reproduction processing circuit 24 in the same manner as the basic data, and FLC data is output.

The FLD circuit 25 performs a fixed length decoding process and outputs it as lower bit data. As shown in FIG. 7, the synthesizing circuit 26 synthesizes the upper bit data and the lower bit data in accordance with the recording mode detected by the mode detecting circuit 21. In the long time mode, all the lower bits are set to 0 and only the upper bits are output. In the standard mode, according to the allocation of the lower bits corresponding to the scaling factor, if the maximum digit of the lower bits is 1, the upper bits are rounded up, so the lower bits are added to the data obtained by subtracting 1 from the upper bits. If the maximum digit of the lower bits is 0, the lower bits are added and output as they are.

This processing eliminates the effect of rounding the high-order bits during recording, and even when playing back a tape recorded in the standard mode in the long-time mode, only the high-order bits that have been rounded are played back. The quantization error is reduced and the image quality is improved as compared with the simple truncation process.

Then, the data output from the synthesizing circuit is inversely quantized by the inverse quantizer 27 and becomes a DCT coefficient.
The 64 DCT coefficients per block are inversely DCT transformed by the IDCT circuit 28, rearranged as image data by the inverse blocking circuit 29, and output from the output terminal 30.

In this embodiment, the image compression recording / reproducing apparatus compatible with both the long time mode and the standard mode is shown. However, even the image compression recording / reproducing apparatus compatible only with the long time mode records in the standard mode. It is possible to play the recorded tape using only basic data.

Further, in the present embodiment, both basic data and additional data have one track per field, but the same effect can be obtained even when recording is performed on a plurality of tracks by changing the rotational speeds of a plurality of recording heads or cylinders. is there.

Further, in this embodiment, the input image data is D
Although the case of performing the CT conversion has been described as an example, it is clear that the image data may be inter-frame difference data, and the same effect can be obtained by any orthogonal conversion such as LOT or Hadamard conversion as the orthogonal conversion.

[0026]

As described above, according to the present invention, the variable length coding is performed on the basic data and the fixed length coding is performed on the additional data according to the property of each data, so that the long time mode and the standard mode can be realized. In either mode, the compression efficiency can be improved.

Since the bit allocation of the lower bits is determined by the characteristics of the adaptive quantizer, it is possible to perform adaptive allocation corresponding to the image data without generating special additional information for that purpose.

Further, in the basic data, the quantizing error can be improved by rounding the reciprocal digit, and in the standard mode, the rounding of the basic data can be restored according to the maximum digit of the lower bit data. The image quality can be improved in any mode.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an image compression recording / reproducing apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing a zigzag scan of DCT coefficients.

FIG. 3 is a diagram showing quantization steps of DCT coefficients.

FIG. 4 is a diagram showing bit allocation of lower bits.

FIG. 5 is a diagram showing bit separation in a separation circuit.

FIG. 6 is a diagram showing a tape pattern.

FIG. 7 is a diagram showing bit combining in a combining circuit.

[Explanation of Codes] 2 Blocking Circuit 3 DCT Circuit 4 Quantizer 5 Separation Circuit 6 VLC Circuit 8 FLC Circuit 11 Mode Change Switch 21 Mode Detection Circuit 23 VLD Circuit 25 FLD Circuit 27 Inverse Quantization Circuit 28 IDCT Circuit 29 Inverse Block Circuit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical indication H04N 7/133 Z

Claims (3)

[Claims] 1. An orthogonal transformer that orthogonally transforms input image data block by block, a quantizer that quantizes output data of the orthogonal transformer, and a quantizer output that is separated into upper bits and lower bits. Separation circuit, variable-length encoder for variable-length encoding the high-order bit output of the separation circuit, first recording head for recording the variable-length encoded data on a recording medium, and low-order bit output of the separation circuit Fixed-length encoder for fixed-length encoding, a second recording head for recording the fixed-length encoded data on the recording medium, and a first reproducing head for reproducing variable-length encoded data from the recording medium. A variable length decoder for variable length decoding the reproduced variable length encoded data, a second reproducing head for reproducing fixed length encoded data from the recording medium, and the reproduced fixed length encoded data Fixed length Fixed-length decoder, a combining circuit that combines data from the variable-length decoder output and the fixed-length decoder output, an inverse quantizer that inversely quantizes the combining circuit output, and the inverse quantizer With an inverse orthogonal transformer that outputs the image data by inverse orthogonal transforming the output of the device, during recording,
In the first recording mode, only the variable length coded data is recorded, and in the second recording mode the variable length coded data,
And both the fixed-length encoded data are recorded, and at the time of reproduction, the synthesizing circuit, when the recording medium is recorded in the first recording mode, the upper bits are the variable-length decoder output, The lower bit is output as 0, and when the recording medium is recorded in the second recording mode, the upper bit is combined with the variable length decoder output and the lower bit is combined with the fixed length decoder output and output. An image compression recording / reproducing apparatus characterized by the above. 2. A quantizer adaptively changes a quantization characteristic in accordance with a characteristic of input image data, and a separation circuit determines a bit allocation of lower bits in accordance with the quantization characteristic. The image compression recording / reproducing apparatus according to claim 1. 3. An orthogonal transformer that orthogonally transforms input image data block by block, a quantizer that quantizes output data of the orthogonal transformer, and a quantizer output that is separated into upper bits and lower bits. Separation circuit, a first recording head for recording the upper bit data of the separation circuit on a recording medium, a second recording head for recording the lower bit data of the separation circuit on the recording medium, and an upper layer than the recording medium A first reproducing head for reproducing bit data, a second reproducing head for reproducing lower bit data from the recording medium, a synthesizing circuit for synthesizing data from the reproduced upper bit data and lower bit data, An inverse quantizer for dequantizing a synthesis circuit output, and an inverse orthogonal transformer for inverse orthogonal transforming the inverse quantizer output to output image data, and And a second recording mode for recording both the high-order bit data and the low-order bit data, the separation circuit rounds and outputs the high-order bits, and the combining circuit In the case of reproducing the second recording mode, when the maximum digit of the lower bit is 1, the image compression recording / reproducing apparatus is characterized in that after subtracting 1 from the upper bit, it is combined and outputted.