JPH07281949A - Data storing method for storage - Google Patents

Abstract

PURPOSE:To reduce an idle area and to effectively store data by setting a control code in the idle area of a storage and then detecting the control code. CONSTITUTION:When the PCM bits are stored in the 8-bit width, it is supposed that the 1st data 1 uses the bit width of the peak value of a series of data. Then lower 8 bits of the data 1 are stored in an address 11 of a memory after the data 1 is regarded as the 10-bit width data, and higher 2 bits of the data 1 are stored in higher 2 bits of an address 12. As the next daa 2 consists of 10 bits, an unused control code 12a is set in an idle area equivalent to 2 bits on the LSB side of the address 12 storing higher bits of the data 2. The data 3 to n-1 can be all stored in the 8-bit width and therefore stored in sequence in the addresses 14-17. The data (n) and its subsequent ones are stored in the 10-bit width and therefore a control code 18a is set in an address 18 after the data n-1 on the address 17 is stored. A write area can be decided by the setting of these control codes.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data storage method for a storage device, and more particularly to a method for efficiently storing data having different bit widths.

[0002]

2. Description of the Related Art At present, when storing data in a storage device, for example, considering the storage device as a memory, it is possible to store more data in the memory without increasing the capacity of the memory. There is a request to do so. In other words, how to effectively use the memory becomes a problem.

On the other hand, the memory is usually accessed in word units, and the number of bits in one word is determined by the memory. Therefore, a problem occurs when the number of bits of data to be stored is different from the number of bits of one word. Figure 3
Will be explained. In FIG. 3, reference numerals 31 to 315 indicate addresses where data is stored, that is, respective words.

For example, consider a case where 10-bit PCM data is stored in a memory in which 8 bits are 1 word, that is, an 8-bit width memory. In this case, 1 data is 1
Since it is 0 bit, a memory area for 2 bytes is always required as shown in FIG. Here, since one word has 8 bits, it is expressed as one byte. More specifically, when data 1 is stored, it is stored in the address 31 of the first byte corresponding to the lower 8 bits. The remaining upper 2 bits are stored in the 2 bits on the MSB side of the address 32 of the second byte. Since the storage area for 6 bits still remains in the address 32, the upper 2 bits of the data 2 which is the next data is stored, and the lower 8 bits of the data 2 are stored in the address 33.
Stores a bit.

In this way, the bits of the address 32 are divided into two, each of which is composed of 4 bits of high-order bits and 4 bits of low-order bits, and stores two different data. That is, the remaining upper 2 bits of the data 1 are stored in 2 bits from the MSB side,
The 4th bit becomes an empty area and stores the code "0",
The upper 2 bits of the data 2 are stored in the 5th and 6th bits, the empty area is stored in the 7th and 8th bits, and the code "0" is stored. Similarly, data after the data 3 is stored. Thus, as shown in FIG. 3, data 1 data n
The +1 storage address is determined.

As described above, in the above-described example, two different data are always stored in the address storing the upper 2 bits, and a 4-bit free area is generated.

In addition, especially for PCM data, 10 bits are not always required. In general, the size of continuous data does not change abruptly, so that the next PCM data also fits in 8 bits or less. In this case, the address data that stores the upper 2 bits of the PCM data is 00H, and this 1 byte is a useless area. That is, when the actual data is represented by 10 bits, a storage area of 3 bytes is required to store two types of data, but when it can be represented by 8 bits or less, it is equivalent to 2 bytes. Since only the storage area is required, a 1-byte useless area is generated.

Image data is another example in which the bit table of input data is different. A method of storing such image data in the memory is shown in FIGS. As described above, this example is a method of storing image data for image display, and is a method of storing image data having different bit widths in one memory. The first image data shown in FIG. 4A is stored in the first memory plane 4a, and is stored between addresses "1" to "5". Next, when the image data is stored as shown in FIG. 4B, "4" pixels are stored with the sixth address as the start address. In this case,
As shown in (b), the first image data is stored.
There is an empty area under the second and second memory planes 4a and 4b. In order to use this free space later for storing new data, free space information as shown in FIG. 5 is generated. FIG. 5 is a list showing the contents of the free area information.
First, the empty area number generated under the image data is No. When it is 1, the leading address is represented by "1", the number of pixels is "5", and the bit width is "2". Next, when storing the third image data as shown in FIG. 4C, the free area information shown in FIG. 5 is searched.
Then, the free area No. 1, the number of pixels and the bit width are compared with the number of pixels and the bit width of the image data,
When it is determined that the image data can be stored, the image data is stored in this empty area.

As described above, in this example, one memory can be efficiently used, but when the image data is packed into the memory later, the empty area information as shown in FIG. 5 is required. Further, this storage method is not suitable for storing PCM data in the memory as shown in FIG. 3 because the memory space is two-dimensional.

[0010]

As described above, in the storage method shown in FIG. 3, first, even if the content of the data represented by 10 bits can be stored in 8 bits, the area for 2 bytes must be stored. Therefore, if the ratio of the data that can be stored in 8 bits is high, the ratio of the useless area becomes high. That is, there is a problem that the storage device has many free areas.

Next, with the method shown in FIGS. 4 and 5, it is possible to solve the problem that there are many free areas as shown in FIG. 3 and use one memory efficiently. However, since the memory space is two-dimensional, it is convenient for storing image data for image display, but it is not suitable for storing PCM data as in the case of the first conventional example. Also, instead of stuffing the data into the memory from the beginning, we are stuffing it later.
In this case, the free area information as shown in FIG. 3 is required, and there is a problem that the information required for storing data increases.

An object of the present invention is to provide an improved data storage method. Especially, it is to provide an optimum method applied to the storage of PMC data.

[0013]

According to a data storage method of a storage device of the present invention, a bit width of write data is stored in a storage device having a predetermined bit width and data having an unspecified bit width is stored in the storage device. And the bit width determined by the storage device are compared, and the write area can be determined by a specific control code.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below with reference to the drawings with reference to the accompanying drawings.

FIG. 1 shows a storage state on a memory of data stored by a method according to an embodiment of the present invention. FIG. 2 is an overall block diagram for realizing the data storage method of the storage device of the present invention including the memory unit for storing the data in FIG.

In FIG. 1, 11 to 111 are addresses where data is stored, and 12a and 18a are specific control codes.

In FIG. 2, reference numeral 21 denotes a buffer, which temporarily holds data to be stored. A detector 22 detects the signal output from the buffer 21 according to a predetermined condition. Based on the detected output, the controller 23 controls the data to change the bit width, and the memory 24 Store the data in.

First, as in the example shown in FIG. 3, 10-bit P
Consider a case where CM data is stored in an 8-bit width memory. First, the buffer 21 stores PCM data of a predetermined one frame. This data is transmitted to both detector 22 and controller 23. Detector 2
In 2, the peak value of the series of transmitted data is detected. The detected information is transmitted to the controller 23, where the bit width of one data to be written is determined. In the case of the present embodiment, the detection information is whether it is within the 8-bit width or larger than the 8-bit width. For example, if the data has a width of 8 bits or less, the detection output 22-1 becomes low level, and if it is larger than that, it becomes high level. In addition,
The first data uses the bit width of the peak value of a series of transmitted data, and is regarded as 10-bit width data in this embodiment. This is because there is no information indicating the actual bit width stored in the memory of the first data.

The controller 23 controls data storage in the memory 24 based on the detected output 22-1. That is, for the first data 1, first, the lower 8 bits of the data 1 are stored in the address 11 of the memory 24, and the upper 2 bits of the data 1 are stored in the upper 2 bits of the address 12, as shown in FIG. Then, the detection output 22-1 from the detector 22 effectively acts on the data 2 and thereafter. Here, the data 2 is data larger than 8 bits (10-bit data), the data 3 to the data n-1 is data that can be stored within 8 bits (8-bit data), and the data n is again stored at 8 bits. There is no data (10-bit data).

Since the data 2 is 10-bit data, the data shown in FIG.
Similar to, store 2 bytes. On the other hand, since data 3 is 8-bit data, it is stored in 1 byte. However, the information indicating that 1 byte has been stored is. Therefore, the controller 23 first sets a control code in the 2-bit free area 12a from the LSB side of the address 12 in which the upper bits of the data 2 are stored. In this embodiment, the control code of the unused area 12a "01" is set. Then, the controller 23 stores the data 3 having an 8-bit width in the address 14 of the memory 24. Since data 4 to data n-1 all fit within an 8-bit width, they are stored in addresses 15 to 17 in order. When reading data, if the control code "01" comes in the area 12a of the address 12, it means that the data is stored in 1 byte from the next data, and the data is read 1 byte at a time. I will

When the controller 23 receives the data n, the detector 22 notifies the output 22-1 that the data cannot be stored in 8 bits. In response, the controller 23 stores in the memory 24 the second control code indicating that the data n is to be stored in 10 bits. That is, after storing the data n-1 of the address 17, the control code 18a is set to the address 18. This code is also referred to as reading information when reading data.
In this embodiment, FFH is set in the control code 18a. That is, data n-1 of address 17 is sequentially
The data is read in 8-bit units, and when FFH appears, it is regarded as a control code, and it means that the data is represented by 10 bits from the next data. Therefore, the data n of the address 19 also reads 2 bits from the MSB side of the address 110. Two bytes of data are read out for one data. When FFH is set, it may be indistinguishable from FFH of 8-bit data. Therefore, FFH of 8-bit data is always represented by 10 bits for distinction. In this way, the data n of the address 19 can be stored in 10 bits.

The storage of the above-mentioned control code and the storage of each data required for the controller 2 can be easily realized by a control circuit and a data generation circuit for normal memory access. For example, it is realized by preparing an 8-bit width register in the controller 23, storing the contents of the address 12 of FIG.

In this embodiment, the storage of two kinds of data of 8 bits and 10 bits is taken up, but this is merely one embodiment, and by setting three or more kinds of control codes, three or more kinds of bits are stored. Width data can be stored, and there is no problem as long as the content of the control code can determine the bit width.

Further, although the type of data and the storage location of the data are PCM data and memory in the present embodiment, there is no problem even if they are different types, and the whole of FIG. It is obvious that there is no problem in the block diagram even if the configuration is different, as long as the data storage as shown in this embodiment can be finally executed.

[0025]

According to the conventional data storage method, there are many free areas, but as described above, the present invention sets the control code in the free area of the storage device and detects the control code. The bit length for storing in the storage device was controlled, and it became possible to store data of different bit widths. As a result, there is an effect that the free area can be reduced and the data can be efficiently stored. For example, the ratio of 8-bit width data in all data is 50%
Assuming that, 16% compression is possible and total storage area is 1/6
You can almost eliminate the empty area.

Further, since the data is packed in the storage device from the beginning, the free area information as in the second conventional example becomes unnecessary, and the storage device can be reduced in size. Further, it is effective when it is desired to store data in a one-dimensional memory space like PCM data.

[Brief description of drawings]

FIG. 1 is a data storage state diagram of a memory stored by a method according to an embodiment of the present invention.

FIG. 2 is an overall system block diagram for performing the method of the present invention.

FIG. 3 is a data storage state of a memory according to a conventional example.

FIG. 4 is a data storage state diagram of a memory according to another conventional example.

FIG. 5 is a diagram showing free area information according to the conventional example of FIG. 4;

[Explanation of symbols]

 11-111 Address where data is stored 12a, 18a Control code 21 Buffer 22 Detector 23 Controller 24 Memory 31-315 Address where data is stored 4a-4d Memory plane

Claims (3)

[Claims] 1. A storage device for storing data having an unspecified bit width in a storage device having a predetermined bit width, comparing the bit width of write data with the bit width determined by the storage device, A data storage method that allows the write area to be identified by a specific control code. 2. It is determined whether or not the supplied data can be represented by the bit width of one address of a memory having a predetermined bit width. If the data can be represented, the first information can be represented. (2) determining means for outputting information, and when the determining means generates the first information, stores data corresponding to the first information in one address of the memory, and when generating the second information, Storing the data corresponding to the above-mentioned data over a plurality of addresses of the memory, and storing the control code indicating the state in the memory each time the output information from the determining means changes, the data storage of the storage device. apparatus. 3. The control code has the same bit width as the bit width of the memory when the first information is changed to the second information, and the control code is changed to the first bit from the second information. The data storage device according to claim 2, wherein a bit width smaller than the bit width is used when the information is changed.