Method for synthesizing speech from text and for spelling all or portions of the text by analogy

• US 5,751,906 A
• Filed: 01/29/1997
• Issued: 05/12/1998
• Est. Priority Date: 03/19/1993
• Status: Expired due to Term

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Nynex Science & Technology
• Inventors: Silverman, Kim Ernest Alexander
• Primary Examiner(s): Hafiz, Tariq R.
• Application Number:
  US08/790,579
• Time in Patent Office: 468 Days
• Field of Search: 395/2.1, 395/2.16, 395/2.17, 395/2.09, 395/2.67-2.7, 395/2.79-2.84
• US Class Current: 704/260
• CPC Class Codes: G10L 13/04 Details of speech synthesis systems, e.g. synthesiser structure or memory management
• CPC Class Codes: G10L 13/08 Text analysis or generation of parameters for speech synthesis out of text, e.g. grapheme to phoneme translation, prosody generation or stress or intonation determination
• CPC Class Codes: G10L 13/10 Prosody rules derived from text; Stress or intonation

Abstract

Improved automated synthesis of human audible speech from text is disclosed. Performance enhancement of the underlying text comprehensibility is obtained through prosodic treatment of the synthesized material, improved speaking rate treatment, and improved methods of spelling words or terms for the sysstem user. Prosodic shaping of text sequences appropriate for the discourse in large groupings of text segments, with prosodic boundaries developed to indicate conceptual units within the text groupings, is implemented in a preferred embodiment.

Claims

1. A method of synthesizing human audible speech from a multi-word string of text, the method comprising the steps of:

treating the multi-word string as a single prosodic paragraph by performing the steps of;

assigning a pitch to the beginning of the multi-word string that is higher than at the end of the multi-word string; and

assigning a pitch to a final end point of the string that is lower than the pitch at any point within the string;

including, in the multi-word string, following at least one of the individual words in the multi-word string, the corresponding spelling of the individual word;

treating each individual word in the multi-word string as a single word declarative sentence;

treating the spelling of each individual word included in the multi-word string as a single word declarative sentence;

grouping each individual word and the corresponding spelling of the individual word into a prosodic group within the single prosodic paragraph, the prosodic group having a higher pitch at the beginning of the prosodic group than at the end of said prosodic group; and

generating speech from the multi-word string as a function of the prosodic groupings and assigned pitch.

2. The method of claim 1, further comprising the steps of:

treating each individual word in the multi-word string as a single word declarative sentence;

treating the spelling of each individual word included in the multi-word string as a single word declarative sentence.

3. The method of claim 1, wherein the spelling of an individual word includes each letter of the individual word.

4. The method of claim 3, further comprising the step of:

inserting, after each letter which is part of the spelling of a word, an additional word beginning with said letter.

5. The method of claim 3, further comprising the step of:

categorizing each letter used to spell an individual word as to whether or not it is to be analogized with another word.

6. The method of claim 5, further comprising the step of:

inserting, after each letter categorized to be analogized to another word which is part of the spelling of a word, an additional word beginning with said letter.

7. The method of claim 6, further comprising the step of selecting the additional word to be inserted following each letter such that it is different from the word being spelled.

8. The method of claim 7, wherein the step of selecting the word to be inserted following each letter involves the step of selecting the word to be inserted from only non-monosyllabic words.

9. The method of claim 7, wherein the step of categorizing each letter as to whether or not it is to be analogized to another word includes the step of:

examining the left and right contexts in which the letter occurs.

10. The method of claim 9, wherein word boundaries are considered when letter contexts are examined.

11. The method of claim 10, further comprising the step of:

arranging successive letters used for the spelling of a word which have been categorized so as not to be analogized with another word into groups; and

inserting a short pause between the groups of letters.

12. A method of synthesizing speech from a segment of text including a first word, comprising the step of:

inserting after the first word, the spelling of the first word; and

generating speech corresponding to the first word and the spelling of the first word.

13. The method of claim 12,

wherein the spelling of the first word includes each letter of the first word;

and

wherein the method further includes the step of inserting after each letter, an additional word beginning with the same letter.

14. The method of claim 12, further comprising the step of:

categorizing each letter used to spell the first word as to whether or not it is to be analogized with another word.

15. The method of claim 14, further comprising the step of:

inserting, after each letter categorized to be analogized to another word, an additional word beginning with said letter.

16. The method of claim 15, further comprising the step of selecting the additional word to be inserted following each letter such that it is different from the word being spelled.

17. The method of claim 16, wherein the step of selecting the word to be inserted following each letter involves the step of selecting the word to be inserted from only non-monosyllabic words.

18. The method of claim 17, wherein the step of categorizing each letter as to whether or not it is to be analogized to another word includes the step of:

examining the left and right contexts in which the letter occurs.

19. The method of claim 18, wherein word boundaries are considered when letter contexts are examined.

20. The method of claim 19, further comprising the step of:

arranging successive letters used for the spelling of a word which have been categorized so as not to be analogized with another word into groups; and

inserting a short pause between the groups of letters.

21. The method of claim 20, further comprising the steps of:

grouping the first word and the spelling of the first word into a prosodic group having a higher pitch at the beginning of the prosodic group than at the end of said prosodic group.

22. The method of claim 21,

wherein the segment of text further includes a second word, the method further comprising the additional step of:

treating the first word, spelling of the first word, and the second word, as a single prosodic paragraph by performing the steps of;

assigning a pitch to the beginning of the first word that is higher than at the end of the second word.