METHOD FOR DIVIDING LETTER SEQUENCES INTO PRONUNCIATION UNITS, METHOD FOR REPRESENTING TONES OF LETTER SEQUENCES USING SAME, AND STORAGE MEDIUM STORING VIDEO DATA REPRESENTING THE TONES OF LETTER SEQUENCES

• US 20150051911A1
• Filed: 04/03/2013
• Published: 02/19/2015
• Est. Priority Date: 04/13/2012
• Status: Abandoned Application

• Current Assignee: Byoung K.I. Choi
• Original Assignee: Byoung K.I. Choi
• Inventors: Choi, Byoung Ki
• Application Number:
  US14/387,841
• Publication Number:
  US 20150051911A1
• Time in Patent Office: Days
• US Class Current: 704/253
• CPC Class Codes: G06F 40/00 Handling natural language data speech analysis or synthesis, speech recognition G10L
• CPC Class Codes: G10L 13/06 Elementary speech units used in speech synthesisers; Concatenation rules
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 25/57 for processing of video signals

Abstract

Disclosed is a method for dividing pronunciation units which includes the steps of: extracting voice-intensity maxima and minima in voice waveforms of letter sequences; forming a group by grouping the extracted maxima together; dividing the letter sequences into pronunciation units around the points nearest to either side of the group from among minima on both sides of the group, voice start points, and voice end points.

Claims

1. A method of dividing a letter sequence into pronunciation units, the method comprising:

extracting maximum points and minimum points of voice intensity from a voice waveform of the letter sequence (S100);

grouping the extracted maximum points to form a group (S200); and

dividing the letter sequence into the pronunciation units, using a point nearest to either side of the group from among the minimum points, a voice start point, and a voice end point as a boundary (S300).

2. The method of claim 1, wherein each pronunciation unit includes one maximum value as a representative value.

3. The method of claim 2, wherein, when a time interval between a specific maximum point and another adjacent maximum point of voice intensity is less than a certain time t1 or when the time interval between the specific maximum point and the other adjacent maximum point of the voice intensity is equal to or greater than the certain time t1 and less than a certain time t2 and a difference between maximum values of the specific maximum point and the other adjacent maximum point is less than a certain level of 1 dB, the grouping of the extracted maximum points (S200) comprises grouping the specific maximum point and the other adjacent maximum point to have a greater value between the maximum values as a representative value.

4. The method of claim 2, wherein, when a time interval between a specific maximum point and another adjacent maximum point of voice intensity is equal to or greater than a certain time t2 or when the time interval between the specific maximum point and the other adjacent maximum point of the voice intensity is equal to or greater than a certain time t1 and less than the certain time t2 and a difference between maximum values of the specific maximum point and the other adjacent maximum point is equal to or greater than a certain level of 1 dB, the grouping of the extracted maximum points (S200) comprises putting the specific maximum point and the other adjacent maximum point in separate groups to have the maximum values of the specific maximum point and the other adjacent maximum point as representative values of the separate groups.

5. A method of representing a tone of a letter sequence, the method comprising:

dividing the letter sequence into pronunciation units;

extracting representative tone data for each of the divided pronunciation units (S400);

calculating tone data for each video frame from the extracted representative tone data and assigning a letter attribute to each video frame (S500); and

playing back the video frame having the letter attribute assigned thereto as a video (S600),

wherein the dividing of the letter sequence into the pronunciation units is performed according to the method of claim 1.

6. The method of claim 5, wherein the representative tone data is voice intensity or voice pitch.

7. The method of claim 6, wherein the representative tone data for the voice intensity includes two boundary points and one maximum point for each pronunciation unit.

8. The method of claim 6, wherein the representative tone data for the voice pitch includes a voice pitch value at a voice start point of the pronunciation unit and a voice pitch value at a voice end point when the voice pitch increases or decreases in the pronunciation unit and includes the voice pitch value at the voice start point, the voice pitch value at the voice end point, and a maximum of maximum values or a minimum of minimum values of the voice pitch in the pronunciation unit when the voice pitch increases and then decreases or decreases and then increases.

9. The method of claim 5, wherein the calculating of the tone data for each video frame from the extracted representative tone data and the assigning of the letter attribute to each video frame (S500) comprises calculating tone data at a time when each video frame is set by interpolation in the representative tone data and then assigning an attribute to a letter in the video frame based on the calculated tone data for each video frame.

10. The method of claim 9, wherein the attribute assigned to the letter includes any one or more of a line thickness, a height, a color, a gradation, a width, a slope, and a size.

11. The method of claim 10, wherein tone data for the voice intensity of the tone data corresponds to a line thickness of a letter and tone data for the voice pitch corresponds to a height of the letter.

12. A method of representing a tone of a letter sequence, the method comprising:

dividing the letter sequence into pronunciation units;

extracting representative tone data for each of the divided pronunciation units (S400);

calculating tone data for each video frame from the extracted representative tone data and assigning a letter attribute to each video frame (S500); and

playing back the video frame having the letter attribute assigned thereto as a video (S600),

wherein the dividing of the letter sequence into the pronunciation units is performed according to the method of claim 2.

13. A method of representing a tone of a letter sequence, the method comprising:

dividing the letter sequence into pronunciation units;

extracting representative tone data for each of the divided pronunciation units (S400);

calculating tone data for each video frame from the extracted representative tone data and assigning a letter attribute to each video frame (S500); and

playing back the video frame having the letter attribute assigned thereto as a video (S600),

wherein the dividing of the letter sequence into the pronunciation units is performed according to the method of claim 3.

14. A method of representing a tone of a letter sequence, the method comprising:

dividing the letter sequence into pronunciation units;

extracting representative tone data for each of the divided pronunciation units (S400);

calculating tone data for each video frame from the extracted representative tone data and assigning a letter attribute to each video frame (S500); and

playing back the video frame having the letter attribute assigned thereto as a video (S600),

wherein the dividing of the letter sequence into the pronunciation units is performed according to the method of claim 4.