Assessment of speech consumability by text analysis

1. A computer implemented method comprising:

obtaining, by one or more processor, an input text, wherein the input text is a script for an output speech;

counting, by the one or more processor, a number of words in a sentence and a number of syllables in the sentence, for each sentence in the input text;

calculating, by the one or more processor, a mean sentence length of the input text as an average number of words in sentences of the input text;

calculating, by the one or more processor, respective variations in sentence lengths for all of the sentences, by subtracting the number of words in each of the sentences from the mean sentence length of the input text;

producing, by the one or more processor, a consumability-readability score of the input text, as an average of the respective variations for all of the sentences in the input text, wherein the consumability-readability score indicates a level of satisfaction for listeners of the output speech based on the input text; and

generating, by the one or more processor, by use of a text-to-speech tool, the output speech based on the input text, based on a preconfigured threshold value for the consumability-readability score of the input text and directives on how to generate the output speech corresponding to a value of the consumability-readability score.

2. The computer implemented method of claim 1 , the generating comprising:

determining that the consumability-readability score of the input text is less than the preconfigured threshold value; and

discarding the input text from candidate scripts for output speeches.

3. The computer implemented method of claim 1 , the generating comprising:

determining that the consumability-readability score of the input text is less than the preconfigured threshold value;

generating a control signal based on the consumability-readability score; and

outputting the control signal to control conversion of the input text to another text semantically similar to the input text in order to improve the consumability-readability score.

4. The computer implemented method of claim 1 , the generating comprising:

determining that the consumability-readability score of the input text is greater than or equal to the preconfigured threshold value; and

generating the output speech based on the input text.

5. The computer implemented method of claim 1 , further comprising:

forming a text matrix for the input text, where rows of the text matrix respectively represent the sentences of the input text in order of appearance, the forming comprising:

instantiating a first column of a row with the number of words in a sentence corresponding to the row;

instantiating a second column of the row with the number of syllables in the sentence corresponding to the row;

instantiating a third column of the row with the number of syllables per word in the sentence corresponding to the row; and

instantiating a fourth column of the row with a difference between the mean sentence length of the input text and the number of words in the sentence corresponding to the row as represented in the first column of the row.

6. The computer implemented method of claim 5 , wherein values in the third column of the text matrix between one (1) and two (2), representing the number of syllables per word in each of the sentences, respectively indicate that each of the sentences represented by each row is easy to read as having relatively short words.

7. The computer implemented method of claim 1 , further comprising:

calculating a conventional readability score of the input text; and

producing the consumability-readability score as multiplied by the conventional readability score in order to reflect a measure of conventional readability in the consumability-readability score.

8. The computer implemented method of claim 7 , wherein the conventional readability score employed is the Flesch Reading Ease Score, denoted as FRES, calculated according to a known formula, and then normalized by percentile, that is, FRES/100.

9. A computer program product comprising:

a computer readable storage medium readable by one or more processor and storing instructions for execution by the one or more processor for performing a method comprising:

obtaining an input text, wherein the input text is a script for an output speech;

counting a number of words in a sentence and a number of syllables in the sentence, for each sentence in the input text;

calculating a mean sentence length of the input text as an average number of words in sentences of the input text;

calculating respective variations in sentence lengths for all of the sentences, by subtracting the number of words in each of the sentences from the mean sentence length of the input text;

producing a consumability-readability score of the input text, as an average of variations for all of the sentences in the input text, wherein the consumability-readability score indicates a level of satisfaction for listeners of the output speech based on the input text; and

updating, by the one or more processor, a text evaluation references database coupled to the one or more processor with a plurality of text evaluation references automatically obtained from data mining and a plurality of consumability-readability scores respectively corresponding to each of the text evaluation references to thereby ascertain a threshold value for the consumability-readability score of the input text in generating the output speech based on the input text.

10. The computer program product of claim 9 , further comprising:

determining that the consumability-readability score of the input text is less than the threshold value; and

discarding the input text from candidate scripts for output speeches.

11. The computer program product of claim 9 , further comprising:

determining that the consumability-readability score of the input text is less than the threshold value;

generating a control signal based on the consumability-readability score; and

outputting the control signal to control conversion of the input text to another text semantically similar to the input text in order to improve the consumability-readability score.

12. The computer program product of claim 9 , further comprising:

determining that the consumability-readability score of the input text is greater than or equal to the threshold value; and

generating the output speech based on the input text.

13. The computer program product of claim 9 , further comprising:

forming a text matrix for the input text, where rows of the text matrix respectively represent the sentences of the input text in order of appearance, the forming comprising:

instantiating a first column of a row with the number of words in a sentence corresponding to the row;

instantiating a second column of the row with the number of syllables in the sentence corresponding to the row;

instantiating a third column of the row with the number of syllables per word in the sentence corresponding to the row; and

instantiating a fourth column of the row with a difference between the mean sentence length of the input text and the number of words in the sentence corresponding to the row as represented in the first column of the row.

14. The computer program product of claim 13 , wherein values in the third column of the text matrix between one (1) and two (2), representing the number of syllables per word in each of the sentences, respectively indicate that each of the sentences represented by each row is easy to read as having relatively short words.

15. The computer program product of claim 9 , further comprising:

calculating a conventional readability score of the input text; and

producing the consumability-readability score as multiplied by the conventional readability score in order to reflect a measure of conventional readability in the consumability-readability score.

16. The computer program product of claim 15 , wherein the conventional readability score employed is the Flesch Reading Ease Score, denoted as FRES, calculated according to a known formula, and then normalized by percentile, that is, FRES/100.

17. A system comprising:

a memory;

one or more processor in communication with memory; and

program instructions executable by the one or more processor via the memory to perform a method comprising:

obtaining an input text, wherein the input text is a script for an output speech;

counting a number of words in a sentence and a number of syllables in the sentence, for each sentence in the input text;

calculating a mean sentence length of the input text as an average number of words in sentences of the input text;

calculating respective variations in sentence lengths for all of the sentences, by subtracting the number of words in each of the sentences from the mean sentence length of the input text;

producing a consumability-readability score of the input text, as an average of variations for all of the sentences in the input text, wherein the consumability-readability score indicates a level of satisfaction for listeners of the output speech based on the input text; and

generating, by use of a text-to-speech tool, the output speech corresponding to the input text, in accordance with directives on how to generate the output speech based on a preconfigured threshold value for the consumability-readability score of the input text.

18. The system of claim 17 , the generating comprising:

determining that the consumability-readability score of the input text is less than the preconfigured threshold value; and

discarding the input text from candidate scripts for output speeches.

19. The system of claim 17 , further the generating comprising:

determining that the consumability-readability score of the input text is less than the preconfigured threshold value;

generating a control signal based on the consumability-readability score; and

outputting the control signal to control conversion of the input text to another text semantically similar to the input text in order to improve the consumability-readability score.

20. The system of claim 17 , the generating comprising:

determining that the consumability-readability score of the input text is greater than or equal to the preconfigured threshold value; and

generating the output speech based on the input text.

21. The system of claim 17 , further comprising:

forming a text matrix for the input text, where rows of the text matrix respectively represent the sentences of the input text in order of appearance, the forming comprising:

instantiating a first column of a row with the number of words in a sentence corresponding to the row;

instantiating a second column of the row with the number of syllables in the sentence corresponding to the row;

instantiating a third column of the row with the number of syllables per word in the sentence corresponding to the row; and

instantiating a fourth column of the row with a difference between the mean sentence length of the input text and the number of words in the sentence corresponding to the row as represented in the first column of the row.

22. The system of claim 17 , further comprising:

calculating a conventional readability score of the input text; and

producing the consumability-readability score as multiplied by the conventional readability score in order to reflect a measure of conventional readability in the consumability-readability score.

23. The computer implemented method of claim 1 , further comprising:

updating a text evaluation references database coupled to the one or more processor with a plurality of text evaluation references automatically obtained from data mining and a plurality of consumability-readability scores respectively corresponding to each of the text evaluation references; and

determining the preconfigured threshold value for the consumability-readability score of the input text to meet to generate the output speech by use of the text-to-speech tool based on the input text.

24. The computer program product of claim 9 , further comprising:

generating, by use of a text-to-speech tool, the output speech based on the input text, based on a the threshold value for the consumability-readability score of the input text and directives on how to generate the output speech corresponding to a value of the consumability-readability score, the generating comprising:

ascertaining that the consumability-readability score of the input test corresponds to discarding the input text from candidate scripts for output speeches, and subsequently discarding the input text;

ascertaining that the consumability-readability score of the input test corresponds to a conversion of the input text to another text semantically similar to the input text in order to improve the consumability-readability score, and subsequently generating the output speech based on a converted input text; and

ascertaining that the consumability-readability score of the input test corresponds to generating the output speech based on the input text, and subsequently generating then output speech based on the input text.

25. The system of claim 17 , further comprising:

updating a text evaluation references database coupled to the one or more processor with a plurality of text evaluation references automatically obtained from data mining and a plurality of consumability-readability scores respectively corresponding to each of the text evaluation references; and

determining the preconfigured threshold value for the consumability-readability score of the input text to meet to generate the output speech by use of the text-to-speech tool based on the input text.