WO2001061543A9 - Method for compression of small computer data files - Google Patents
Method for compression of small computer data filesInfo
- Publication number
- WO2001061543A9 WO2001061543A9 PCT/US2001/004986 US0104986W WO0161543A9 WO 2001061543 A9 WO2001061543 A9 WO 2001061543A9 US 0104986 W US0104986 W US 0104986W WO 0161543 A9 WO0161543 A9 WO 0161543A9
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- data
- compression
- entries
- ofthe
- storing
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
-
- H—ELECTRICITY
- H03—ELECTRONIC CIRCUITRY
- H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
- H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
- H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
Definitions
- the present invention generally relates to compression of binary data for ease of transmission between a source and receiving device, and more particularly to compression of binary data files by selecting an optimum compression method from a plurality of compression methods.
- the definition ofthe codes is large relative to the data to be compressed, and second, when data changes slightly from record to record, few repetitive byte values can be assigned compact codes to represent them.
- the present invention although useful for many applications, was originally designed for the purpose of reducing file sizes of data recorded at remote locations of natural
- the present invention has reduced the size ofthe file needed to be transmitted to a
- the present invention eliminates unnecessary bits and makes use ofthe principle that
- Computers store and use numbers in binary format (base 2), instead ofthe more familiar base 10 format. Computers typically use a 16-bit or 32-bit segment of binary
- Binary digits, or bits may have a value of
- 1024 (base 10) is expressed as 00000100 00000000 (base 2 or binary). 1023 is expressed as 00000011 11111111, and uses one less meaningful bit than the value of 1024. Just as in base 10 notation, the zeros included in the 16-bit number before the first digit that is not zero do not have any numerical significance.
- the present invention examines data arrays found in gas flow records and
- each of five algorithms is applied to each array and the method is chosen for each array that allows maximum compression.
- Each compression method makes
- the present invention will use only four bits instead of the 16 bits or 32 bits used in the standard computer format. Many ofthe compression methods and techniques are in
- present invention also includes novel methods for specifying time and summary data.
- Some data arrays such as flow time, may include many repetitions ofthe same
- the present invention searches for the most frequently repeated value in the data
- the present invention will include the mode only one time in the compression record. If the array has variations from the mode, the variations are treated as outliers and treated individually. The present invention identifies the position of outlier
- Mode compression Some data arrays, such as static pressure or C readings, may change by small
- Some data arrays such as temperature readings or storage tank level readings, increase or decrease regularly by relatively small amounts.
- the present invention will include
- This compression method is called
- Some data arrays are proportional to another array or a combination of other
- volume is proportional to C, pressure
- the present invention creates an array of estimated values and finds a ratio that when multiplied to the estimated values will minimize the range of deltas between the multiplied
- Ratio compression This compression method is called Ratio compression.
- 100,000 is expressed in binary form as 00000000 00000001 10000110 10100000.
- the present invention can safely strip the first 15 bits from the binary representation, as they will be 0 for every value contained in the array.
- Strip Zeros This compression method is called Strip Zeros. Each ofthe five compression methods also makes use of other compression techniques to further reduce the size ofthe compressed file. Strip Top Bit is routinely used to remove the first bit of compressed values when it is safe to assume that the value ofthe top
- the outlier compression method allows the program to identify which elements of an array are “outliers" or expressed differently than the majority of other
- an outlier bitmap is a series of bits that "maps" which values of an array must be treated as outliers.
- bit of an array of data entries when that bit may be assumed to be there.
- Fig. 1 is a top-level processing diagram, or flowchart, showing the process of the present invention
- Fig. 2a is a table illustrating a sample data array
- Fig. 2b is a compression record created for the data contained in Fig. 2a;
- Fig. 3a is a table illustrating a sample data array
- Fig. 3b is a compression record created for the data contained in Fig. 3a;
- Fig. 4 is a compression record illustrating the Strip Top Bit compression
- Fig. 5 is a top-level processing diagram, or flowchart, showing the process of
- Fig. 6a is a table illustrating a sample data array
- Fig. 6b is a compression record created for the data contained in Fig. 6a
- Fig. 7a is a table illustrating a sample data array with binary values
- Fig. 7b is a compression record created for the data contained in Fig. 7a;
- Fig. 7c is another compression record for the data contained in Fig. 7a;
- Fig. 7d is another compression record for the data contained in Fig. 7a;
- Fig. 8 is a top-level processing diagram, or flowchart, showing the process of
- Fig. 9 is a top-level processing diagram, or flowchart, showing the process of
- Fig. 10a is a table illustrating a sample data array
- Fig. 10b is a compression record for the data contained in Fig. 10a;
- Fig. 1 la is a table illustrating a sample data array
- Fig. 1 lb is a compression record for the data contained in Fig. 1 la
- Fig. 12a is a table illustrating a sample data array
- Fig. 12b is a compression record for the data contained in Fig. 12a;
- Fig. 13a is a table illustrating a sample data array
- Fig. 13b is a compression record for the data contained in Fig. 13a;
- Fig. 14 is a top-level processing diagram, or flowchart, showing the process of the Difference From Last compression method;
- Fig. 15a is a table illustrating a sample data array
- Fig. 15b is a compression record for the data contained in Fig. 15a;
- Fig. 16a is a table illustrating a sample data array
- Fig. 16b is a compression record for the data contained in Fig. 16a;
- Fig. 17a a table illustrating a sample data array
- Fig. 17b a compression record for the data contained in Fig.17a;
- Fig. 18 is a top-level processing diagram, or flowchart, showing the process of
- Fig. 19 is a table illustrating a sample data array
- Figs. 20a - 20d are tables showing iterations ofthe mantissa determination of
- Fig. 21a is a table illustrating a sample data array
- Fig. 21b is a table showing estimate data used by the Ratio compression
- Fig. 21c is a table showing deltas determined by the Ratio compression
- Fig. 21d is a compression record for the data contained in Fig. 21a;
- Fig. 22a is a table illustrating a sample data array
- Fig. 22b is a table showing estimates determined by the Ratio compression method
- Fig. 22c is a compression record for the data contained in Fig. 22a;
- Fig. 23a is a table illustrating a sample data array
- Fig. 23b is a table illustrating an intermediate step ofthe Ratio compression
- Fig. 23c is a compression record for the data contained in Fig. 23a.
- the present invention was designed to compress the binary representation of numbers. It performs its compression by removing unnecessary bits from binary
- 00000576 identifies the same value as 576.
- 00000100 00000000 in binary format identifies the same value as 100 00000000.
- N bits 2 N - 1.
- 4 bits can hold a maximum value of 2 4 - 1 or 15.
- 15 is represented in binary form as l l l l.
- base 10 the maximum value of N
- N digits 10 N - 1.
- 4 digits in base 10 can hold a maximum value of 10 4 - 1 or 9999.
- Fig. 1 a flowchart illustrates the basic steps included within the present invention.
- Raw data is input.
- the present invention immediately compresses time data (time of day that readings were taken) and removes "no
- Time data readings are typically one hour apart at the top ofthe hour. Under
- records where flow did not occur are identified when all readings are 0 except for temperature and static temperature. These records are removed from the data to be compressed, and indices for which records are no flow are included in the compressed file.
- each of five algorithms are applied to each data array and the best method is selected. The five
- the present invention then saves the data in data fields limited to that number of bits,
- Mode Compression searches every parameter array for the value most frequently repeated, or the mode value. For example, when flow time is consistently 60 minutes, Mode Compression will store one value along with an instruction to repeat that value in each field ofthe array. Any deviations from the mode are identified with the Outlier Compression. If there are no exceptions, the present invention will only save the beginning time value and need not include the delta standard
- the Delta From Base method finds the lowest value in the array and subtracts that
- the Difference From Last method is similar to Delta From Base, but the deltas are the differences from the previous values instead of from a base number. This method works particularly well for parameters that change relatively uniformly over time, such as temperature and tank levels. Finally the Ratio Compression method is used when the
- elements of an array are related proportionally to a data set that can be created from other arrays already included in the compression record.
- the present invention finally performs a
- Strip Top Bit Each technique minimizes the number of bits required to identify the data to
- Outlier values are those in data arrays that require unique identification, allowing the remainder ofthe elements to be expressed in a more compact format.
- Fig. 2a shows an array of 24 data entries having 21 identical entries and three
- the varying entries being the first, third, and fourth locations.
- bitmap indicates with a 1 in the first, third, and fourth locations ofthe 24-element bitmap that the first, third, and fourth data entries are outliers. If the present invention were to use this configuration, it would require 24 bits to specify which elements were outliers, plus an additional bit to specify that more than one outlier exists. The value of the 21 identical entries is included only once. Although the outlier values must be stored individually, the
- Fig. 2b shows a compression record where that same identification is accomplished with only 13 bits. As shown in Fig. 2b, the first bit indicates
- the number of outliers indicator is safely reduced by one because at least one outlier must
- the next bit indicates whether the index is recorded in reverse order from the actual order ofthe data array. In situations where the outliers are towards the
- the present invention examines both directions ofa data array and
- the final index can only be one less than 2, so the final index
- the outlier method can be used to determine whether outliers are congregated at one end of an array.
- Fig. 3a shows a 24-element data array
- elements 15, 17 and 20 - 24 are outliers. Because five outliers are clustered at the end ofthe array, the last four positions do not have
- Fig. 3a The compression record for Fig. 3a is shown in Fig. 3b.
- the invention must know the lengths ofthe fields.
- the present invention will size the fields to match the largest size
- the present invention may determine that the first bit of every data entry is 1. For example, in a Delta From Base array, the base number might be
- Strip Top Bit must deal with 0 as an element. It does this by setting the field width value to maximum. For example, when compressing an array with a maximum value of 2 30 - 1, five bits are necessary to specify the field width. If the largest value is in fact 2 30 - 1, then the file size will be stored as 11110. To express 0, the field width will be set to the maximum size or 11111. In other words, 0 is specified as 2 N - 1 , where N is
- the present invention in its optimal form, works only with positive integer values.
- the upper limit is bound by Strip Top Bit, which uses the value 31 to specify
- This number limits bit widths for data to 30, which has a maximum decimal amount of 1,073,741,823. To ensure error free execution, the present invention may adjust the raw data to remain below this maximum number.
- the present invention adds an offset to all temperature data to make certain that the data consists only of positive integers.
- all temperature data By default, all temperature
- the present invention checks all the data
- offset results in -1000, which is less than 0. 0 is substituted for -1000.
- the present invention could also address negative and positive numbers by adding a sign entry where necessary, either as an inflection indicator or a sign indicator for each data entry.
- Strip Zeros The first compression method analyzed by the present invention is Strip Zeros.
- the present invention recognizes that 15 of these bytes are not necessary and only the last 17
- Strip Zeros saves 15 bits for a single data entry less the bits necessary to
- Strip Zeros evaluates four distinct methods of compressing arrays of numbers
- the four methods are: (1) uniform field size; (2) single
- Strip Zeros is selected as
- the present invention determines that all array elements are exactly N bits in size. Only N - 1 bits are stored because the present
- Fig. 6a shows a 6-index array
- Fig. 6b shows the compression record for that array, demonstrating how the uniform size indicator ofthe
- Strip Zeros compression method was used and how each element was identified with 6 bits.
- each data entry was reduced by a size of 5 bits.
- the single field size Strip Zeros compression method is used when the uniform field size does not apply. In other words, some values will require different size fields than others.
- the bit width ofthe storage field is selected to be the bit width necessary to describe the largest element. Any preceding zeros are stripped from the values, but the top
- the individual field sizes method can take advantage ofthe Strip Top Bit technique since each element is separately compressed. Before each element in the compression record is a field size
- the maximum size ofthe field can be predetermined.
- the method can save storage space.
- the original record was 76 bits, while the compressed record
- the two field sizes Strip Zeros compression method divides the array elements into two groups.
- the present invention starts with the first group being limited to data values
- the present invention may determine that the smallest data value requires 4 bits to identify. All data values that can be identified in 4 bits will belong in the
- the size ofthe compressed file is calculated. Strip Zeros next increases the bit width ofthe smaller group by one, so that the small group includes the previous data entries as well as all entries identifiable with one more bit. The present invention then recalculates the results. This process is repeated until the bit width ofthe smaller group is
- Fig. 7d shows a compression record
- the result in this case is 66 bits. While that size causes the present invention to
- the present invention may be able to save an additional bit by listing the larger field width before the smaller field width. For example, if
- the Mode Compression method searches through an array for the value most
- a reading may be taken every hour, resulting in precisely 60 minutes between readings.
- the present invention identifies the value ofthe data, strips the top bit, and transmits the value one time instead of once for each element.
- the number of array elements is either a fixed number or specified at the beginning
- Mode Compression adds an outlier map and the values ofthe outlier elements
- Mode Compression is capable of reducing the data file by the number of bits necessary to store all data elements minus one (the one that must be transmitted) less the bits necessary to define the compression. Decompression occurs simply by copying the saved
- the Delta From Base compression method finds the lowest value in the array.
- This value is called the base and is subtracted from all the array elements to create deltas.
- Delta From Base yield particularly good compression when the base is significantly greater than the largest delta. For example, suppose an array had values of 6000, 6003, 6015, and
- the present invention can send the
- Delta From Base employs a number of strategies for adjusting the base, specifying the deltas and determining the outliers to minimize the overall compression record
- Delta From Base begins with the consideration of one delta bit width and no outliers.
- An example of an array that Delta From Base compresses efficiently is shown in
- the base bit width, after Strip Top Bit is applied, is 8 bits.
- the largest delta, 4022, requires 1 ' 2 bits to identify.
- the base is transmitted one time in 8 bits and the
- Bin size refers to the field width necessary to identify the data ofthe array
- Delta From Base performs this analysis by starting with the largest
- element value as the base. For example, examine the data array shown in Fig. 11a. The largest value, selected for the base, is 30803. The deltas range from 0 to 493. The largest delta thus requires 9 bits to identify. The greatest compression ratio occurs, however, when
- Fig. 1 lb shows the compression record for this array and this method, indicating
- the base width requirement is 14.
- the top bit is stripped, and the larger delta of 493 is assigned its width of 9 bits.
- the smaller delta size is assigned a width of 4 bits.
- This example has small bin deltas clustered at the end ofthe array, so the compression record need not include indices for all of them.
- An outlier is defined as a delta that is too
- Delta From Base can express the outlier as a raw value or as a value outside the delta range.
- the goal behind removing outliers is to shrink the number of bits required to identify the remaining records that vary little from the base number, and thereby reduce the overall number of bits in the compressed file. For example, suppose that in an array of 24
- Delta From Base adjusts the base to the new value, saves the new upper and lower delta values, and determines the field width necessary to handle the base and the
- maximum delta would be 730 and would require 10 bits to store. If three values were
- Delta From Base can use a separate size field of six bits to hold 33 and 39. Another approach is to subtract the smaller delta range of 33, leaving 0 and 6, which fit in only 3 bits.
- This approach has the advantage of using fewer bits for specifying outliers, but does take an
- Fig. 13a shows a 24-element array with a delta range from 0 to
- the outlier value is shifted down by 512 (2 9 ) to 31, which takes 5 bits to store.
- the adjusted base is represented only once and removed from the remaining elements.
- the outlier element has an additional
- deltas are the differences from the previous values and not from a base value. This compression type works particularly well for parameters that steadily ramp up/down or cycle through a sampling period, such as a tank level or temperature sensor. Difference From Last
- the values of 6000, 6003, and 6015 have binary values of: 00010111 01110000;
- Difference From Last may report the second and third values as 0011 and l l l l, saving the first 12 bits of each entry. Decompression requires that the first delta be added to or
- map It consists of a number of bits equivalent to the number of data entries in the data array.
- the first method splits the differences into a large size bin and a small size bin.
- Fig. 15b shows the compression record for the array.
- Fig. 16a shows an array where the initial value is 4547 and the differences range from -1131 to 994. Difference From Last will create an inflection record to identify when the differences change between positive and negative values. The inflection record specifies where the differences
- the record contains an inflections indicator, few outliers indicator, number of outliers, index direction backwards indicator, and the outlier
- index records. It then indicates the initial size ofthe field to hold the initial value and the
- Fig. 17a shows an
- Difference From Last sets all the differences to positive. Difference From Last starts with 10 bits to store 739 and then shrinks the difference size from 10 to 0 bits. In this example, the optimal size occurs at 5 bits, leaving three outliers in index
- Ratio compression is used when the elements of an array are related
- Processing begins by creating an array of estimated values, and then
- the present invention searches for the optimum way to describe the deltas.
- ratio compression For compressing natural gas flow data, ratio compression is used for three types of parameters: pressure extension, volume and energy. For each, an estimated array is
- the estimated arrays are created in the following ways:
- the estimated pressure extension array is calculated by taking the square
- the estimated volume array is created by multiplying flow time by pressure
- the pressure extension array is not included in the gas flow data record, then the estimated pressure extension array is used. If either the flow time or C array
- the estimated energy array is simply the actual volume array.
- the estimated arrays are created in a similar manner from
- the present invention checks to make sure that multiplying the differential and static pressures will not result in an overflow before performing the square root, i.e., that the result will not exceed the maximum storage of 32 bits.
- the product size is estimated by
- log 2 (product) ⁇ log 2 (differential pressure) + log 2 (static pressure).
- Log 2 is defined as the number of bits required to store a value. If the product were to overflow, then log 2 (product) would exceed 32 by no more than N bits where
- N log 2 (differential pressure) + log 2 (static pressure) - 32. If the present
- the present invention shifts down the larger ofthe two elements first until it is the smaller, and then alternates shifting down the respective elements. This
- the present invention prevents an estimated volume overflow with
- estimated volume pressure extension x flow time x C.
- the present invention performs the calculation in two steps: first multiplying the pressure extension by flow time and then multiplying the result by C. Either step can result in an overflow, defined in this case as exceeding the maximum Strip Top Bit representation of a number, meaning 30 bits.
- the product is re-written in logarithmic form as previously described and Shrink Product is used to avoid overflows. In this case, the two equations are rewritten log 2 (product) ⁇
- log 2 (multiplicand 1) + log 2 (multiplicand2). If the product were to overflow, then log 2 (product) would exceed 30 by N, where N log 2 (multiplicand 1) + log 2 (multiplicand2) - 30,
- the present invention will correct this with an adjusting ratio.
- the data arrays may be amenable for Ratio compression. If an array is proportionally related to another, a simple constant multiplied to one array will provide the values of another. In this situation, the original array may be used as the estimate array. In the situation where the second array is
- ratios may include two offsets that vary over time. In this case, the best approach is to use the
- Ratio compression assumes that the ideal ratio is the one that minimizes the sum of deltas when specified as:
- the present invention uses a more compact representation by limiting the mantissa to 15 bits (14 bits after
- the ratio record can range from 1 bit to 24 bits. Allowing the mantissa to become greater than 15 bits
- the dynamic range ofthe ratio is from 4.66 x 10 "10 to 3.28 x 10 4 .
- the present invention calculates the initial ratio in four steps:
- the present invention aborts further ratios processing if either array sum is 0, the calculated
- the present invention will also take steps to prevent an overflow in summing ratio arrays. Before summing either the estimate or actual arrays, the present invention will verify that the result will not exceed the maximum storage size of 32 bits. The sum is quickly estimated as:
- the present invention increases the exponent by N when preventing an estimated array sum
- the table presented in Fig. 19 is a sample gas flow record with 22 volume
- the estimated volume was created by multiplying flow time, C and the square root ofthe product of differential and static pressures, as previously
- the present invention checks for a potential overflow in summing the actual and estimated volume arrays.
- an overflow may exist.
- volume sum 15016.
- the present invention adjusts the estimated and actual volume
- the present invention divides the actual volume sum (3936354304) by the estimated volume sum (184651) to create the mantissa (21317).
- the present invention recognizes that the exponent is greater than 31, and that it needs make no further adjustments.
- the present invention must then determine the compression record size corresponding to this ratio. Next, it will shrink the mantissa by one
- the present invention saves these values.
- the present invention treats the new ratio as a better approximation ofthe optimum ratio.
- Fig. 21a For example, consider the pressure extension array of Fig. 21a.
- the present invention creates an estimate array by taking the square root ofthe product ofthe differential and static pressures as shown in Fig. 21b.
- the present invention determines that the optimum ratio has a mantissa
- the present invention picks the lowest delta as the base and adjusts the deltas upwards. Since
- Ratio Compression next considers dividing the deltas into two size bins. Ratio searches every possible lower bin size, which can range from 0 to one less than the largest bin size. Ratio repeats this process by using the largest delta as the base and subtracts each delta from it. If any two-size combination from either search results in an overall smaller compression record, ratio compression will save the parameters. For example, see Fig. 22a.
- the present invention determines that the optimum ratio has a mantissa of 3 and an exponent
- ratio crates a range of deltas from -455 to 15 as seen in Fig. 22b.
- the present invention picks
- delta bin size or field bit width, one bit at a time and then calculates the value ofthe difference between the top ofthe data range and the outliers. Ratio repeats this process by using the largest delta as the base and subtracting each delta from it. If any combination
- Fig. 23a shows an
- the present invention uses the square root ofthe product ofthe differential and static pressures for the pressure extension estimate.
- the present invention determines that the optimum ratio has a mantissa of 1 and an exponent of 0. Since the ratio is 1 , the present invention uses a single flag to indicate a ratio of one in the
- FIG. 23b shows an array for the same example readings as Fig. 23a with
- Fig. 23c shows the compression record for this example. The first 5 lines ofthe record indicate that the data is compressed by ratio compression with many outliers and indicate the positions of
- the base is indicated in the next three lines as positive 68.
- a flag indicating direction from base for the outliers is next, bit lengths are specified, and the outlier values are
- the present invention applies Time Compression, No Flow Record Removal, Daily Summary Compression, and One
- the present invention can compress the
- the present invention creates an outlier bitmap for all the data entries.
- the first entry is flagged as an outlier. Any other entry that is not exactly 60 minutes from the preceding entry is also flagged as an outlier.
- the full value of all outliers is stored in the compressed file. During decompression, the outlier values are
- the invention searches the data for records for "no flow" records. No flow records exist when all readings are zero with the exception of static pressure and temperature. The invention then removes the zero elements form the
- present invention uses the outlier method to identify which records are no flow.
- the present invention creates a summary value of each given parameter array after compression.
- the summary value may be the daily average or it may be a total across
- the present invention begins by calculating the average or total of a
- the present invention first checks for an overflow, i.e., the sum exceeds the maximum storage space of 32 bits.
- the upper limit ofthe sum is estimated by:
- log 2 (sum) ⁇ log 2 (largest element) + log 2 (number of elements). If the sum were to overflow, then log 2 (sum) would exceed 32 by no more than N bits
- N log 2 (largest element) + log 2 (number of elements) - 32.
- the present invention determines that N is greater than 0, it can prevent an overflow by shifting down each element N bits before summing and shifting the average back up by N bits.
- the present invention sets the total to the upper limit. Otherwise, it checks if the [( ⁇ elements) x 24] portion ofthe extrapolation will exceed 32 bits by calculating
- N log 2 (largest element) + log 2 (24) - 32. If N is greater than 0, the present invention can prevent an overflow by first shifting down ⁇
- total ( ⁇ [ ( £ elements) / 2 N ] x 24 ⁇ / number of elements) x 2 N .
- the present invention determines the difference with the actual summary. Depending on the size ofthe difference,
- the present invention will create on of three types of compression records. If the difference is not
- the small size summary compression record is used.
- the compression record includes a small size indicator (1 bit), a negative sign indicator (1 bit) and a value for
- the present invention can assume the size ofthe summary value in the compression record.
- the compression record will include a flag for the summary size being other than small size, a flag to indicate the value is within the calculated range, and the summary value.
- the compression record must add a designator for the size ofthe field
Abstract
Description
Claims
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2001238365A AU2001238365A1 (en) | 2000-02-15 | 2001-02-15 | Method for compression of small computer data files |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US50415300A | 2000-02-15 | 2000-02-15 | |
US09/504,153 | 2000-02-15 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2001061543A1 WO2001061543A1 (en) | 2001-08-23 |
WO2001061543A9 true WO2001061543A9 (en) | 2002-10-17 |
Family
ID=24005072
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/US2001/004986 WO2001061543A1 (en) | 2000-02-15 | 2001-02-15 | Method for compression of small computer data files |
Country Status (2)
Country | Link |
---|---|
AU (1) | AU2001238365A1 (en) |
WO (1) | WO2001061543A1 (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1292036B1 (en) * | 2001-08-23 | 2012-08-01 | Nippon Telegraph And Telephone Corporation | Digital signal decoding methods and apparatuses |
US6577254B2 (en) * | 2001-11-14 | 2003-06-10 | Hewlett-Packard Development Company, L.P. | Data compression/decompression system |
US7672262B2 (en) | 2005-04-22 | 2010-03-02 | Baker Hughes Incorporated | System, method, and apparatus for command and control of remote instrumentation |
EP2153610A4 (en) * | 2007-06-01 | 2010-09-15 | Research In Motion Ltd | Interactive compression with multiple units of compression state information |
GB2471056B (en) * | 2010-03-09 | 2011-02-16 | Quantum Corp | Controlling configurable variable data reduction |
CN107409152B (en) * | 2015-03-12 | 2021-10-15 | 英特尔公司 | Method and apparatus for compressing data received over a network |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
SE8307228D0 (en) * | 1983-12-30 | 1983-12-30 | Grundstenen 16808 Ab | data compression |
US4988998A (en) * | 1989-09-05 | 1991-01-29 | Storage Technology Corporation | Data compression system for successively applying at least two data compression methods to an input data stream |
US5649151A (en) * | 1992-06-29 | 1997-07-15 | Apple Computer, Inc. | Efficient method and apparatus for access and storage of compressed data |
US5546575A (en) * | 1994-05-23 | 1996-08-13 | Basil E. Potter & Associates, Inc. | Encoding method for compressing a tabular database by selecting effective compression routines for each field and structure of partitions of equal sized records |
US5678043A (en) * | 1994-09-23 | 1997-10-14 | The Regents Of The University Of Michigan | Data compression and encryption system and method representing records as differences between sorted domain ordinals that represent field values |
-
2001
- 2001-02-15 WO PCT/US2001/004986 patent/WO2001061543A1/en active Application Filing
- 2001-02-15 AU AU2001238365A patent/AU2001238365A1/en not_active Abandoned
Also Published As
Publication number | Publication date |
---|---|
WO2001061543A1 (en) | 2001-08-23 |
AU2001238365A1 (en) | 2001-08-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CA2475186C (en) | Method and apparatus for windowing in entropy encoding | |
US8037034B2 (en) | Methods of creating a dictionary for data compression | |
Lewis | Interleaved memory function interpolators with application to an accurate LNS arithmetic unit | |
CA2283591C (en) | Data coding network | |
US8452825B2 (en) | Sortable floating point numbers | |
CN101095284B (en) | Device and data method for selective compression and decompression and data format for compressed data | |
EP1995974B1 (en) | Method for realizing arithmetic coding | |
US20070050436A1 (en) | Order-preserving encoding formats of floating-point decimal numbers for efficient value comparison | |
US7973680B2 (en) | Method and system for creating an in-memory physical dictionary for data compression | |
CN104834539B (en) | A kind of data increment update method | |
US20110234432A1 (en) | Systems and methods for optimizing bit utilization in data encoding | |
US6253222B1 (en) | Compression of limited range floating point numbers | |
US5995970A (en) | Method and apparatus for geographic coordinate data storage | |
WO1996021200A1 (en) | Method and apparatus for digital data compression | |
CN107291832B (en) | Data storage method based on list storage structure | |
US7921144B2 (en) | Fast correctly-rounding floating-point conversion | |
US5724253A (en) | System and method for searching data vectors such as genomes for specified template vector | |
WO2001061543A9 (en) | Method for compression of small computer data files | |
US6240431B1 (en) | Decompression of limited range floating point numbers | |
CN104077272A (en) | Method and device for compressing dictionary | |
KR102243119B1 (en) | Quantization apparatus and method using numerical representation with dynamic precision | |
Chandrasekaran et al. | Polynomial testing of the query “Is ab≥ cd?” with application to finding a minimal cost reliability ratio spanning tree | |
JP3018990B2 (en) | Arithmetic coding device | |
US20240005213A1 (en) | System and method for off-chip data compression and decompression for machine learning networks | |
JPH0137047B2 (en) |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG UZ VN YU ZA ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
AK | Designated states |
Kind code of ref document: C2 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG UZ VN YU ZA ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: C2 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG |
|
COP | Corrected version of pamphlet |
Free format text: PAGES 1/30-30/30, DRAWINGS, REPLACED BY NEW PAGES 1/33-33/33; DUE TO LATE TRANSMITTAL BY THE RECEIVING OFFICE |
|
REG | Reference to national code |
Ref country code: DE Ref legal event code: 8642 Ref country code: DE Ref legal event code: 8642 |
|
122 | Ep: pct application non-entry in european phase | ||
NENP | Non-entry into the national phase in: |
Ref country code: JP |