Method of identifying duplicate voice recording

• US 7,571,093 B1
• Filed: 08/17/2006
• Issued: 08/04/2009
• Est. Priority Date: 08/17/2006
• Status: Active Grant

• Current Assignee: National Security Agency
• Original Assignee: The United States of America As Represented By The Director National Security Agency
• Inventors: Cusmariu, Adolf
• Primary Examiner(s): Sked; Matthew J
• Application Number:
  US11/506,090
• Time in Patent Office: 1,083 Days
• Field of Search: None
• US Class Current: 704/207
• CPC Class Codes: G10L 25/48 specially adapted for particular use
• CPC Class Codes: G10L 25/90 Pitch determination of speech signals

Abstract

A method of identifying duplicate voice recording by receiving digital voice recordings, selecting one of the recordings; segmenting the selected recording, extracting a pitch value per segment, estimating a total time that voice appears in the recording, removing pitch values that are less than and equal to a user-definable value, identifying unique pitch values, determining the frequency of occurrence of the unique pitch values, normalizing the frequencies of occurrence, determining an average pitch value, determining the distribution percentiles of the frequencies of occurrence, returning to the second step if additional recordings are to be processed, otherwise comparing the total voice time, average pitch value, and distribution percentiles for each recording processed, and declaring the recordings duplicates that compared to within a user-definable threshold for total voice time, average pitch value, and distribution percentiles.

Claims

1. A method of identifying duplicate voice recording, comprising the steps of:

a) receiving a plurality of digital voice recordings;

b) selecting one of said plurality of digital voice recordings;

c) segmenting the selected digital voice recording;

d) extracting a pitch value from each segment;

e) estimating a total time that voice appears in the selected digital voice recording;

f) removing pitch values that are less than and equal to a user-definable value;

g) identifying unique pitch values in the result of step (f);

h) determining the frequency of occurrence of the unique pitch values;

i) normalizing the result of step (h) so that the frequencies of occurrence are greater than zero and less than one;

j) determining an average pitch value from the pitch values remaining after step (f);

k) determining the distribution percentiles of the result of step (h);

l) if additional digital voice recordings are to be processed then returning to step (b), otherwise proceeding to the next step;

m) comparing the results of steps (e), (j), and (k) for each digital voice recording processed; and

n) declaring the digital voice recordings duplicates that compared to within a user-definable threshold for each of the results of steps (e), (j), and (k).

2. The method of claim 1, wherein the step of receiving a plurality of digital voice recordings is comprised of the step of receiving a plurality of digital voice recordings in any digital format.

3. The method of claim 2, wherein the step of segmenting the selected digital voice recording is comprised of the step of segmenting the selected digital voice recording into 16 millisecond segments sampled at 8000 samples per second.

4. The method of claim 3, wherein the step of extracting a pitch value from each segment is comprised of the step of extracting a pitch value from each segment using any pitch extraction method.

5. The method of claim 4, wherein the step of estimating a total time that voice appears in the selected digital voice recording is comprised of the step of estimating a total time that voice appears in the selected digital voice recording using the pitch values.

6. The method of claim 5, wherein the step of removing pitch values that are less than and equal to a user-definable value is comprised of the step of removing pitch values that are less than and equal to zero.

7. The method of claim 6, further including the step of removing pitch values that vary from one pitch value to the next pitch value by less than or equal to a user-definable value.

8. The method of claim 7, wherein the step of normalizing the result of step (h) so that the frequencies of occurrence are greater than zero and less than one is comprised of the step of dividing the result of step (h) by the number of pitch values remaining after step (f).

9. The method of claim 8, wherein the step of determining an average pitch value from the pitch values remaining after step (f) is comprised of the step of determining an average pitch value from the pitch values remaining after step (f) and rounding to the nearest integer.

10. The method of claim 1, wherein the step of segmenting the selected digital voice recording is comprised of the step of segmenting the selected digital voice recording into 16 millisecond segments sampled at 8000 samples per second.

11. The method of claim 1, wherein the step of extracting a pitch value from each segment is comprised of the step of extracting a pitch value from each segment using any pitch extraction method.

12. The method of claim 1, wherein the step of extracting a pitch value from each segment is comprised of the step of extracting a pitch value from each segment using a cepstral pitch extraction method.

13. The method of claim 1, wherein the step of estimating a total time that voice appears in the selected digital voice recording is comprised of the step of estimating a total time that voice appears in the selected digital voice recording using the pitch values.

14. The method of claim 1, wherein the step of removing pitch values that are less than and equal to a user-definable value is comprised of the step of removing pitch values that are less than and equal to zero.

15. The method of claim 1, further including the step of removing pitch values that vary from one pitch value to the next pitch value by less than or equal to a user-definable value.

16. The method of claim 1, wherein the step of normalizing the result of step (h) so that the frequencies of occurrence are greater than zero and less than one is comprised of the step of dividing the result of step (h) by the number of pitch values remaining after step (f).

17. The method of claim 1, wherein the step of determining an average pitch value from the pitch values remaining after step (f) is comprised of the step of determining an average pitch value from the pitch values remaining after step (f) and rounding to the nearest integer.