Method for compressing data

• US 4,890,326 A
• Filed: 03/03/1988
• Issued: 12/26/1989
• Est. Priority Date: 03/03/1988
• Status: Expired due to Fees

• Current Assignee: Rubiyat Software Incorporated
• Original Assignee: Rubiyat Software Incorporated
• Inventors: Euler, James A.
• Primary Examiner(s): Shoop, Jr., William M.
• Assistant Examiner(s): Romano, Gary J.
• Application Number:
  US07/163,658
• Time in Patent Office: 663 Days
• Field of Search: 341/51, 341/56, 341/63, 341/67, 364/135, 364/821, 381/30, 381/31, 381/35
• US Class Current: 704/229
• CPC Class Codes: G10L 19/00 Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis in musical instruments G10H
• CPC Class Codes: H03M 7/30 Compression speech analysis-synthesis for redundancy reduction G10L19/00; for image communication H04N; Expansion; Suppression of unnecessary data, e.g. redundancy reduction

Abstract

A method for compressing and decompressing voice data enables efficient voice storage on small computers. Analog voice data is converted to digital voice data and the difference jumps between adjacent numbers in the digital voice data are measured. In the preferred embodiment, if the value of the jump is within the range +2 to -2, then a code value is assigned to that jump from zero to four where the code value equals the jump value plus two. If the jump value is outside the range, a jump is normally assigned a code value of five. Three adjacent codes are compressed to one code using the formula:

Compression number=6×[6×(code 3)+ code 2] +code 1

and at least this compression number is stored. If one or two of the code values in a group of three code values has a value of five, the actual jump value is stored after the compression number. However, if all jumps in a group of three have an absolute value of greater than two and less than twenty-three, then a second compression option is chosen in which a compression word is generated that will identify the three jump values, and the compression word is stored as part of the compressed data. After the aforementioned compression steps, further compression is performed to eliminate repetitious zeroed jumps. Decompression is accomplished by reading the compressed data, recognizing the compression numbers and compression words and determining the codes and jumps from the compression numbers, the compression words, and the compressed data itself.

Claims

1. A method for compressing and decompressing analog voice data for storage in a computer memory comprising:

converting the analog voice data to digital voice data comprising a series of digital numbers;

determining the difference jumps between adjacent numbers in the digital voice data;

assigning a code value to each jump;

grouping the codes into groups of adjacent codes and calculating a single compression number for each group that is unique for any combination of code values in the group;

storing at least the compression numbers as compressed data; and

reading the compressed data and selectively utilizing the stored compression numbers to determine codes of each group.

2. A method for compressing and decompressing analog voice data for storage in a computer memory comprising:

converting the analog voice data to digital voice data comprising a series of digital numbers;

determining the difference jumps between adjacent numbers in the digital voice data;

assigning a code value to each jump where at least some of the code values equal the jump value plus M, M being a selected integer;

grouping the codes into groups of three adjacent codes and calculating a single compression number for each group according to the formula;

space="preserve" listing-type="equation">Compression number=N[N(Code

     3)+Code 2]+Code 1

where N=2 (M+1) and Codes 1, 2 and 3 are three adjacent code values of a group;

storing at least the compression numbers as compressed data; and

reading the compressed data and selectively utilizing the stored compression numbers to determine codes 1, 2 and 3 of each group.

3. The method of claim 2 further comprising:

assigning a code value of 2M+1 to each jump having an absolute value of greater than M;

for each jump having an absolute value greater than N, storing in the compression data after the compression number the actual jump value;

after the compression number has been utilized to determine codes 1, 2 and 3, determining the jump values from each code having a value of less than 2M plus one and reading the compression data stored after the compression number to determine the value of the jumps corresponding to each code in a group having a value of 2M+1.

4. The method of claim 2 wherein said reading and utilizing step further comprises:

dividing the compression number by N to determine a first whole number quotient and a first remainder;

defining the first remainder as code 1;

dividing the first whole number quotient by N to find a second whole number quotient and a second remainder;

defining the second remainder as code 2;

defining the second whole number quotient as code 3.

5. The method of claim 2 wherein M equals two.

6. The method of claim 2 further comprising:

if all three jumps in a group of three have an absolute value of greater than M, generating a compression word as determined by the following formula;

space="preserve" listing-type="equation">Compression word=K[K(Code

     3)+Code 2]+Code 1

where the code values are assigned according to the following formulas;

space="preserve" listing-type="equation">Code value=Jump Value+K/2-M

for positive jumps and

space="preserve" listing-type="equation">Code value=Jump Value=K/2+M

for negative jumps;

storing a flag number in the compression data to indicate the presence of the compression word in the data;

storing the compression word after the flag number;

recognizing the flag number to indicate the presence of a compression word following the flag number;

reading the next two bytes after the flag number as the compression word;

dividing the compression word by K to determine a first whole quotient and a first remainder;

defining the first remainder as code 1;

dividing the first whole quotient by K to determine a second whole quotient and a second remainder;

defining the second remainder as code 2; and

defining the second whole quotient as code 3.

7. The method of claim 6 where K equals forty and M equals two.

8. The method of claim 2 further comprising:

storing the compression number Q when all three jumps in a group of three are equal to zero;

scanning the stored data to locate compression numbers having a value of Q;

counting the number of adjacent compression numbers having a value of Q;

when two or more adjacent compression numbers have a value of Q, replacing the first of said compression numbers with a value of R to serve as a flag;

deleting from the compression data the adjacent compression numbers having a value of Q; and

storing data after the R value in the compression data to specify the count of consecutive Q values that were encountered in the compression data.