JP4749896B2 - Image display device - Google Patents

Description

  The present invention relates to an image display device that converts image data in a block data format into image data in a raster data format and stores the image data in a display refresh memory.

For example, in the JPEG (Joint Photographic Experts Group) compression method, discrete cosine transform (DCT: Discrete Cosine Transform) is used when compressing image data.
In the DCT compression technique, image data is divided into blocks of 8 pixels × 8 pixels and the data (block data format data) is processed.
However, it is necessary to continuously input image data in line units to the display monitor receiver. That is, it is necessary to continuously input image data to the display monitor receiver in the screen scanning order (scanning from left to right and scanning from top to bottom with respect to one entire image). is there.
Therefore, it is necessary to convert the image data in the block data form into the image data in the raster data form, and supply the image data in the raster data form to the display monitor receiver. A buffer memory is needed.

Until new image data cannot be stored in the buffer memory until all the image data is read from the buffer memory in which the image data for 8 lines is stored, 8 lines are required to continuously process the image data. A buffer memory capable of storing a minute amount of image data is further required.
As described above, the conventional image display device prepares two systems of 8-line buffer memories, and writes the block data form image data to one 8-line buffer memory while the other 8-line buffer memory is being written. The image data stored in is read in the form of raster data.
Since the image data is composed of a luminance component Y and color components U and V, two lines of 8-line buffer memory are required for each component (see, for example, Patent Document 1).

The processing contents of the conventional image display apparatus will be specifically described below.
For example, it is assumed that a JPEG decoder constituting an image display device outputs image data of a pixel group (MCU: Minimum Coded Unit) of 16 pixels × 8 pixels in the form of block data. FIG. 3 is an explanatory diagram showing the configuration of an MCU composed of 16 pixels × 8 pixels.
When the sub-sampling specification is YUV = 4: 2: 2, the block data form image data is component data (Y data (luminance data), U data (color data), V data) constituting image data of one MCU. (Color data)) is data that is output in a unit of 256 bytes in the order of Y data (128 bytes) → U data (64 bytes) → V data (64 bytes). FIG. 4 is an explanatory diagram showing the output order of image data in the form of block data.

When converting three component data (block data of Y data, U data, and V data) constituting one MCU into image data in a raster data format (display data format) that is output dot-sequentially, As shown in FIG. 5, a buffer memory (data rearrangement memory) corresponding to the number of pixels for one horizontal scanning effective period × the number of block configuration lines (8 lines) × luminance / color data byte size is prepared.
Then, the block data is temporarily stored in the buffer memory, and the block data is converted into dot-sequential raster data image data corresponding to display horizontal / vertical scanning, and the raster data image data is displayed. Store in refresh memory.

However, new block data is written to the buffer memory and updated until the output of the block data of 640 columns × 8 rows written in the buffer memory as image data in the form of raster data from the buffer memory is completed. I can't.
For this reason, the JPEG decoder can output block data continuously for one frame at an update rate of 1 byte / clock, but after writing block data of 640 column pixels × 8 rows of pixels, 640 column pixels × 8 The operation is to wait until the block pixel block data is output as raster data image data. Similarly, the output of image data in the form of raster data is an operation of waiting for the output until the writing of the block data of 640 columns × 8 rows is completed, and the output speed is halved.

Therefore, as shown in FIG. 6, the conventional image display apparatus stores 640 columns × 8 rows × 2 banks (system) in order to perform the conversion operation without waiting time and minimize the output period. A buffer memory (data rearranging memory) that can be used is prepared, and a configuration is adopted in which writing and reading are alternately switched for two buffer memories.
As shown in FIG. 7, the processing for writing the image data in the raster data format into the display refresh memory is as follows. The raster data format image data is raster data in the form of a continuous pixel data stream for one horizontal scanning effective period. Since it is designed on the assumption that there is a buffer memory (data for the number of pixels for one horizontal scanning effective period (640 pixels) x the number of block configuration lines (8 lines) x luminance / color data byte size (2 bytes) The rearrangement memory is provided with a plurality of systems (two systems). That is, a 20480-byte buffer memory (data rearrangement memory) is provided.

JP 2004-159330 A (pages 7 to 8, FIG. 1)

  Since the conventional image display device is configured as described above, it is designed on the assumption that image data in the form of raster data is raster data in the form of a continuous pixel data stream for one horizontal scanning effective period. . Therefore, in order to avoid halving the output speed, it is necessary to prepare a large number of buffer memories (data rearranging memories), and there is a problem that the cost is increased.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide an image display apparatus that can avoid half of the output speed without preparing a large number of buffer memories.

In the image display device according to the present invention, the image data conversion means is a variable for specifying the order in which block data arrives, the number of horizontal MCUs, the block data identification number, the number of vertical pixels in the block data, and the horizontal Using the number of directional pixels, the memory write address corresponding to the raster data output order is calculated, and the block data is written to the memory write address in the buffer memory on which the raster data is not read by the mapping means. By executing the processing, the raster data is stored in the buffer memory, and the mapping means reads the raster data from the buffer memory that has not been written by the image data conversion means, and the horizontal direction in the display refresh memory From start address to end address Then, when the raster data is sequentially written to the addresses that are continuous in the horizontal direction and the writing of the raster data to the end address is finished, in the line one line down in the vertical direction, from the horizontal start address to the end address, Raster data is sequentially written at addresses that are continuous in the horizontal direction.

According to this invention, the image data conversion means is a variable that specifies the order of arrival of block data, the number of horizontal MCUs, the block data identification number, the number of vertical pixels in the block data, and the number of horizontal pixels. Is used to calculate the memory write address corresponding to the output order of the raster data, and the block data write processing is performed at the memory write address in the buffer memory that is not subjected to the raster data read processing by the mapping means. Thus, the raster data is stored in the buffer memory, and the mapping means reads the raster data from the buffer memory that has not been written by the image data conversion means, and starts from the horizontal start address in the display refresh memory. Horizontal to the end address When the raster data is written sequentially to the consecutive addresses, and the raster data is written up to the end address, it continues in the horizontal direction from the top address to the end address in the horizontal direction on the next line in the vertical direction. Since the raster data is sequentially written to the addresses to be written , there is an effect that half of the output speed can be avoided without preparing many buffer memories.

Embodiment 1 FIG.
FIG. 1 is an explanatory diagram showing a configuration and data conversion operation of an image display device according to Embodiment 1 of the present invention. In FIG. 1, a JPEG decoder 1 is a pixel group (MCU: Minimum Coded Unit) composed of a plurality of pixels. ) Image data, that is, MCU image data compressed by the JPEG compression method, and the MCU image data is grouped for each of the luminance component Y and the color difference components U and V of the pixel image data ( (Hereinafter referred to as block data). The JPEG decoder 1 constitutes decoding means.
The image data converter 2 performs a process of converting the block data output from the JPEG decoder 1 into raster data image data (hereinafter referred to as raster data) corresponding to the dot sequence of the pixels constituting the MCU. The image data conversion unit 2 constitutes image data conversion means.

The buffer memory 3 is a data rearrangement memory that temporarily stores raster data converted by the image data conversion unit 2.
The mapping unit 4 performs a process of storing the raster data stored in the buffer memory 3 in the SDRAM 5 so that the memory mapping of the raster data corresponds to the horizontal / vertical scanning for display. The buffer memory 3 and the mapping unit 4 constitute mapping means.
The SDRAM 5 is a display refresh memory that stores raster data.

Next, the operation will be described.
When the JPEG decoder 1 receives MCU image data compressed by the JPEG compression method, the JPEG decoder 1 collects the MCU image data for each of the pixel luminance component Y and the color difference components U and V, and outputs block data.
That is, when the data ratio YUV of the Y data (luminance data), U data (color data), and V data (color data) constituting the image data of the MCU is 4: 2: 2, the JPEG decoder 1 is shown in FIG. In the order of addresses 00 (hexadecimal) → FF (hexadecimal) shown in Fig. 8, Y data (128 bytes), U data (64 bytes), and V data (64 bytes) are collected, and block data of a total of 256 bytes is converted into MCU. Output every time.
As shown in FIG. 3, one MCU occupies a screen of 16 pixels × 8 lines when corresponding to the display screen, and constitutes one display screen of VGA size with horizontal MCU 40 × vertical 60 MCU. Is.

When the image data conversion unit 2 receives block data from the JPEG decoder 1, the image data conversion unit 2 performs a process of converting the block data into raster data corresponding to the dot sequence of the pixels constituting the MCU.
That is, the image data conversion unit 2 performs dot-sequential scanning (U · Y · V · Y...) If the mapping unit 4 reads the data from the buffer memory 3 while incrementing the read address by 1 from 0. Based on the arrival order of the block data output from the JPEG decoder 1, the write address for the buffer memory 3 is calculated.

Each time the block data is received from the JPEG decoder 1, the image data converter 2 calculates a write address for the buffer memory 3. When data is written to the write address, the raster data is stored in the buffer memory 3 as a result. Will be.
In the first embodiment, two systems of 256-byte buffer memory 3 are prepared. The reason is that the block operation data output from the JPEG decoder 1 can be continuously and alternately written and read out without waiting for the block data to maximize the conversion operation speed.

  In the first embodiment, the image data conversion unit 2 converts block data into raster data, stores the raster data in the buffer memory 3, and the mapping unit 4 reads the raster data stored in the buffer memory 3. The image data conversion unit 2 stores block data in the buffer memory 3 and the mapping unit 4 converts the block data into raster data when reading out the block data stored in the buffer memory 3. You may do it.

Hereinafter, the process in which the image data conversion unit 2 converts block data into raster data will be specifically described.
First, the order of arrival of block data is the number of MCUs MH in the horizontal direction, the identification number G of the data block (block 0_Y, block 1_Y, block 0, 1_U, block 0, 1_V), the number of vertical pixels L in the data block, the data block Is expressed by the following equation (1) using the horizontal pixel number P as a variable.
256MH + 64G + 8L + P = 0 to 511 bytes (1)
However,
MH: 0 to 1 ... 1MH = 4G
G: 0-3 ... 1G = 8L
0 → block0_Y
1 → block1_Y
2 → block0,1_U
3 → block0, 1_V
L: 0-7 ... 1L = 8P
P: 0 to 7

Next, the output order of raster data for scanning the pixel configuration of the MCU is as follows: the number of MCUs in the horizontal direction MH, the number of vertical pixels L in the data block, and the YUV in the MCU (Y: U: V = 4: 2: It is expressed by the following formula (2) using the data number N of 2) cycle number W and 1 cycle YUV (Y: U: V = 4: 2: 2) as a variable.
256MH + 32L + 4W + N = 0 to 511 bytes (2)
However,
MH: 0 to 1 ... 1MH = 8L
L: 0-7 ... 1L = 8W
(1 MCU is composed of 8 rows of ITU-R.656 format data = 8 rows of 32 bytes of data)
W: 0 to 7 (16 pixels / MCU / 2 pixels / cycle) 1W = 4N
N: 0 to 3
0 → U data 1 → Y even data 2 → V data 3 → Y odd data

The image data conversion unit 2 counts the order of the block data output from the JPEG decoder 1 and generates variables [MH, G, L, P] constituting the above equation (1).
Then, the image data conversion unit 2 derives variables [MH, L, W, N] constituting the equation (2) from the variables constituting the equation (1), and raster data for scanning the pixel configuration of the MCU. Memory write address ADRS [8. . 0] is generated.
ADRS [8..0] = 256MH [0] + 32L [2..0] + 4W [2..0] + N [1..0];
case G is
when 0 => W [2..0] = P [2..1] +0; N [1..0] = (P [0], 1);
when 1 => W [2..0] = P [2..1] +4; N [1..0] = (P [0], 1);
when 2 => W [2..0] = P [2..0]; N [1..0] = 0;
when 3 => W [2..0] = P [2..0]; N [1..0] = 2;
end case;

  As described above, the image data converter 2 reads the memory write address ADRS [8. . 0] is generated, its address ADRS [8. . 0], the raster data is stored in the buffer memory 3 by executing the block data writing process.

When the image data conversion unit 2 stores the raster data in the buffer memory 3, the mapping unit 4 converts the raster data stored in the buffer memory 3 so that the memory mapping of the raster data corresponds to the horizontal / vertical scanning for display. A process of storing in the SDRAM 5 is performed.
Hereinafter, the writing process of the mapping unit 4 to the SDRAM 5 will be specifically described.

In the writing process of the mapping unit 4, first, the head address ofstx (0 × 8 + 0 to 7) +0 is calculated according to the generation of the synchronization timings sync_FP, sync_MV, sync_MCU, sync_MH shown in FIG.
When the start address ofstx (0 × 8 + 0 to 7) +0 is calculated, the operation of storing 16 pixels of data (32 bytes) in successive addresses is repeated for 8 lines, and then the start address ofstx (0 X8 + 0-7) + 1x32 is calculated.
When the head address ofstx (0 × 8 + 0 to 7) + 1 × 32 is calculated, the operation of storing data for 16 pixels (32 bytes) at successive addresses is repeated for 8 lines.
Next, the head address ofstx (0 × 8 + 0 to 7) + 2 × 32 is calculated, and the operation of storing 16 pixels of data (32 bytes) at successive addresses is repeated for 8 lines.

As described above, the mapping unit 4 evaluates the number of occurrences of the synchronization timings sync_FP, sync_MV, sync_MCU, and sync_MH shown in FIG. 2, so that the counting variables MV (0 to 59), MCU (0 to 39), MH (0 ~ 7) is generated, the head address ofstx (MV × 8 + MH) + MCU × 32 is calculated, and the operation of storing 16 pixels of data (32 bytes) at successive addresses is repeated for 8 lines. Execute.
Thus, for each MCU, the operation is performed for 40 MCUs in the horizontal direction × 60 MCUs in the vertical direction, and the raster data corresponding to one display screen of VGA size is stored in the SDRAM 5, while the raster data from the SDRAM 5 is synchronized with the display timing. By reading, display output data in the form of raster data in the screen scanning order is obtained.

  As is apparent from the above, according to the first embodiment, the block data output from the JPEG decoder 1 is converted into raster data corresponding to the dot sequence of the pixels constituting the MCU, and the raster data is converted into the raster data. An image data conversion unit 2 to be stored in the buffer memory 3 is provided, and the mapping unit 4 stores the raster data stored in the buffer memory 3 in the SDRAM 5 so that the memory mapping of the raster data corresponds to the horizontal / vertical scanning for display. Thus, the output speed can be halved without preparing a large number of buffer memories.

  That is, according to the first embodiment, the time for holding the block data output from the JPEG decoder 1 is minimized as much as possible by using only a total of 4,096 bits of buffer memory 3 of 1 MCU (256 bytes) × 2 systems. Thus, by generating raster data in MCU units and applying a display memory control function according to the raster data, the required memory capacity of the buffer memory 3 is reduced from the conventional 163,840 bits (20,480 bytes) to 4,096. The cost can be reduced by reducing the number of bits to 512 bytes.

It is explanatory drawing which shows the structure and data conversion operation | movement of an image display apparatus by Embodiment 1 of this invention. It is explanatory drawing which shows the write-in process of the raster data with respect to SDRAM. It is explanatory drawing which shows the structure of MCU which consists of 16 pixels x 8 pixels. It is explanatory drawing which shows the output order of the image data of a block data form. It is explanatory drawing which shows the structure and data conversion operation | movement of the conventional image display apparatus. It is explanatory drawing which shows the structure and data conversion operation | movement of the conventional image display apparatus. It is explanatory drawing which shows the structure and data conversion operation | movement of the conventional image display apparatus.

Explanation of symbols

  1 JPEG decoder (decoding unit), 2 image data conversion unit (image data conversion unit), 3 buffer memory (mapping unit), 4 mapping unit (mapping unit), 5 SDRAM (refresh memory for display).

Claims (3)

Inputting image data of the MCU is a group of pixels and a plurality of pixels, collectively image data of the MCU for each luminance and chrominance of the pixel, outputs the block data is image data of the block data forms Decoding means, image data conversion means for converting the block data output from the decoding means into image data in a raster data format corresponding to the dot sequence of the pixels constituting the MCU , and the image data conversion means A raster memory that temporarily stores raster data , which is image data in the form of raster data converted by, and raster data stored in the buffer memory are read out, and the memory mapping of the raster data is used for horizontal and vertical scanning for display. map for storing the raster data so as to correspond to the display refresh memory and For example Bei and packaging means,
Two systems of buffer memory are prepared,
The image data converting means calculates the number of the MCUs in the horizontal direction, the identification number of the block data, the number of vertical pixels and the number of horizontal pixels in the block data, which are variables specifying the order in which the block data arrives. The memory write address corresponding to the output order of the raster data is calculated, and the block data write process is performed on the memory write address in the buffer memory that is not subjected to the raster data read process by the mapping means. To store the raster data in the buffer memory,
The mapping means reads raster data from the buffer memory that has not been written by the image data conversion means, and continues in the horizontal direction from the horizontal start address to the end address in the display refresh memory. When the raster data is sequentially written to the address to be written and the raster data is written to the end address, the raster data continues in the horizontal direction from the top address to the end address in the horizontal direction on the next line in the vertical direction. An image display device, wherein the raster data is sequentially written to an address.   2. The image display device according to claim 1, wherein the decoding means inputs image data of a pixel group compressed by the JPEG compression method.   In the decoding means, the data ratio of luminance data, U color data, and V color data constituting the image data of the pixel group is 4 to 2: 2, and the number of pixels constituting the pixel group is 16 pixels in the horizontal direction. 3. The image display device according to claim 1, wherein the number of pixels is eight in the vertical direction.

Images