Method and device for speech analysis

Abstract

A device and a method for speech analysis are provided, comprising measuring fundamental notes of a speech sequence to be analysed and identifying frequency intervals between at least some of said fundamental notes. An assessment is then made as to the frequency at which at least some of these thus identified intervals occur in the speech sequence to be analysed. Among other applications are speech training and diagnosis of pathological conditions.

1. (C) A method for analysing the human speech, comprising:
 (A) measuring sounding fundamental notes of a speech sequence to be analysed; 
 (B) identifying the frequency intervals between at least two of the fundamental notes in succession; and 
 (C) assessing the frequency at which at least one of the frequency intervals thus identified occurs in the speech sequence to be analysed wherein the measurement of fundamental notes in the speech sequence to be analysed comprises distinguishing fundamental notes of sufficient duration above a predetermined threshold value and separating fundamental notes of sufficient loudness, whereby only the fundamental notes thus separated are identified. 
 
     
     
       2. A method as claimed in  claim 1 , wherein the measurement of fundamental notes in the speech sequence to be analysed comprises separating fundamental notes of sufficient loudness above a predetermined amplitude threshold value. 
     
     
       3. A method of speech synthesizing, comprising the steps of analysing at least one speech sequence from at least one person by means of an analysing method as defined in  claim 2 , and on the basis of said the analysis controlling the production of synthetic speech on the basis of at least one aspect of said the analysis. 
     
     
       4. A method as claimed in  claim 3 , wherein said the analysis comprises an appraisal of several speech sequences from the same individual. 
     
     
       5. A method as claimed in  claim 3  or  4 , wherein said the analysis comprises an appraisal of speech sequences from several different individuals. 
     
     
       6. A method as claimed in  claim 1 , wherein measuring the fundamental notes in the speech sequence to be analysed comprises establishing a pitch mean value during a predetermined time interval, and using the mean value thus obtained for identification of fundamental notes. 
     
     
       7. A method as claimed in  claim 1 , wherein identifying the frequency intervals also comprises identification of whether the frequency interval is ascending or descending. 
     
     
       8. A method as claimed in  claim 1 , wherein at least minor thirds (L 3 ) and major thirds (S 3 ) are identified. 
     
     
       9. A method as claimed in  claim 1 , comprising identification of at least one frequency interval from the group: unison (R 1 ), minor second (L 2 ), major second (S 2 ), minor third (L 3 ), major third (S 3 ), minor sixth (L 6 ), and or major sixth (S 6 ); and at least one frequency interval from the group perfect fourth (R 4 ), augmented fourth/diminished fifth (tritone), perfect fifth (R 5 ), minor seventh (L 7 ), and or major seventh (S 7 ). 
     
     
       10. A method as claimed in  claim 1 , wherein the identified frequency intervals are divided into at least the following sub-groups: unison (R 1 ), minor second (L 2 ), major second (S 2 ), minor third (L 3 ), major third (S 3 ), minor sixth (L 6 ), and or major sixth (S 6 ); respectively: perfect fourth (R 4 ), augmented fourth/diminished fifth (tritone), perfect fifth (R 5 ), minor seventh (L 7 ), major seventh (S 7 ), and or perfect octave (R 8 ). 
     
     
       11. A method as claimed in  claim 1 , comprising identifying frequency interval sequences between notes of at least one group comprising at least three of the fundamental notes in succession. 
     
     
       12. A method as claimed in  claim 11 , wherein the identified frequency interval sequences comprise notes of a major or minor chord. 
     
     
       13. A method as claimed in  claim 11 , wherein the identified frequency interval sequences comprise ascending or descending arpeggios of major or minor chords. 
     
     
       14. A device for analysing human speech, comprising:
 (A) measurement means for measuring sounding fundamental notes in a speech sequence to be analysed comprising a filtering unit for separation of fundamental notes having a duration above a predetermined time threshold value and for separating fundamental notes of a loudness above a predetermined amplitude threshold value; 
 (B) means for identifying frequency intervals between at least two of these fundamental notes in succession; and 
 (C) means for estimating the frequency at which at least one of the frequency intervals thus identified occurs in the speech sequence to the analysed. 
 
     
     
       15. A device as claimed in  claim 14 , wherein the measurement means for measuring fundamental notes further comprises means for establishing the pitch mean value during a predetermined time interval. 
     
     
       16. A device as claimed in  claim 14 , wherein the means for identification of frequency intervals is designed to at least identify minor thirds (L 3 ) and or major thirds (S 3 ). 
     
     
       17. A device as claimed in  claim 14 , wherein the means for identification of frequency intervals further is designed to identify the frequency interval sequences between notes of at least one group comprising at least three of the fundamental notes in succession. 
     
     
       18. A device for synthetic speech, comprising a device for analysing at least one speech sequence from at least one individual in accordance with  claim 14 , and means for producing synthetic speech, said the means for producing synthetic speech being controlled on the basis of at least some aspect of the analysis produced by said the analysing device. 
     
     
       19. A method for automated speech training, comprising:
 (A) providing an uttered speech sequence by a user; 
 (B) measuring sounding fundamental notes of the speech sequence to be analysed; 
 (C) identifying frequency intervals between at least two of these fundamental notes in succession; 
 (D) assessing the frequency at which at least one of the frequency intervals thus identified occurs in the speech sequence to be analysed; and 
 (E) comparing the estimated frequency interval frequency with a preferred frequency determined in advance for the user concerned. 
 
     
     
       20. A method as claimed in  claim 19 , said the method further comprising presenting the results of the comparison between the assessed frequency interval frequency and the preferred frequency determined in advance for the user concerned. 
     
     
       21. A method as claimed in  claim 19 , said the method further comprising identifying suitable measures for reduction of the difference between the assessed frequency interval frequency and the preferred frequency determined in advance. 
     
     
       22. A method as claimed in  claim 19 , wherein the method is carried out in real time. 
     
     
       23. A method as claimed in  claim 19 , wherein the preferred frequency determined in advance for the user concerned comprises standard values. 
     
     
       24. A method as claimed in  claim 23 , wherein said the standard values are grouped into at least one of the categories user types and or speech training purpose. 
     
     
       25. A device for speech training, comprising:
 (A) means for registering a spoken speech sequence; 
 (B) means for measuring sounding fundamental notes of the speech sequence to be analysed; 
 (C) means for identifying frequency intervals between at least two of these fundamental notes in succession; 
 (D) means for assessing the frequency at which at least one of the frequency intervals thus identified occurs in the speech sequence to be analysed; and 
 (E) means for comparing the assessed frequency interval frequency with a preferred frequency determined in advance for the user concerned. 
 
     
     
       26. A device as claimed in  claim 25 , said the device further comprising means for presenting the results of the comparison between the assessed frequency interval frequency and the preferred frequency determined in advance for the user concerned. 
     
     
       27. A device as claimed in  claim 25  or  26 , said the device further comprising means for identifying suitable measures for reduction of the difference between the assessed frequency interval frequency and the preferred frequency determined in advance. 
     
     
       28. A device as claimed in  claim 25 , said the device further comprising a database having at least one set of standard values to be used as the preferred frequency determined in advance for the user concerned. 
     
     
       29. A device as claimed in  claim 28 , said the device further comprising a plurality of sets of standard values grouped with regard to at least one of the categories user types and speech training purpose. 
     
     
       30. A method for diagnosing pathological conditions on the basis of a speech analysis, comprising:
 (A) uttering a speech sequence by a patient; 
 (B) measuring sounding fundamental notes of the speech sequence to be analysed; 
 (C) identifying frequency intervals between at least two of these fundamental notes in succession; 
 (D) assessing the frequency at which at least one of the frequency intervals thus identified appears in the speech sequence to the analysed; and 
 (E) evaluating at least one such assessed frequency interval frequency by comparing it with a frequency determined in advance for diagnostic purposes. 
 
     
     
       31. A method as claimed in  claim 30 , wherein the frequency determined in advance is based on at least one corresponding previous analysis of a speech sequence from the same patient. 
     
     
       32. A method as claimed in  claim 31 , wherein the frequency determined in advance is based on appraisal of corresponding analyses of at least two speech sequences from the same patient. 
     
     
       33. A method as claimed in  claim 31 , wherein the frequency determined in advance is based on appraisal of corresponding analyses of more than two speech sequences from the same patient. 
     
     
       34. A method as claimed in  claim 30 , wherein the frequency determined in advance is based on normal values. 
     
     
       35. A method as claimed in  claim 34 , wherein patients are further grouped in categories, the frequency determined in advance being based on the normal values applying to the category of users concerned. 
     
     
       36. A method as claimed in  claim 34 , said the method further comprising presenting the results of the evaluation of the comparison between the frequency interval frequency and the frequency determined in advance. 
     
     
       37. A device for diagnosing pathological conditions on the basis of a speech analysis, comprising:
 (A) means for registration of an uttered speech sequence; 
 (B) measurement means for measuring sounding fundamental notes of the speech sequence to be analysed; 
 (C) identification means for identifying frequency intervals between at least two of these fundamental notes in succession; 
 (D) assessment means for assessing the frequency at which at least one of the frequency intervals thus identified appears in the speech sequence to be analysed; and 
 (E) means for evaluating at least one such assessed frequency interval frequency by comparing it with a frequency determined in advance for diagnostic purposes. 
 
     
     
       38. A device as claimed in  claim 37 , said the device further comprising presentation means for presenting the assessed frequency interval frequency. 
     
     
       39. A device as claimed in  claim 37  or  38 , said the device further comprising a database having at least one set of standard values to be used as the preferred frequency determined in advance for the user concerned, and preferably a plurality of sets of standard values grouped with regard to at least one of the categories user types and diagnosing purpose. 
     
     
       40. A device as claimed in  claim 37  or  38 , said the device further comprising a plurality of sets of standard values grouped with regard to at least one of the categories user types and diagnosing purpose. 
     
     
       41. A method for speech analysis, the method comprising:
 identifying a plurality of notes of a speech sequence of an individual;   identifying frequency intervals between at least two of the plurality of notes;   determining a frequency at which at least one of the identified frequency intervals occurs in the speech sequence;   comparing the frequency of the at least one identified interval frequency with a second frequency; and   determining a characteristic of the individual based at least on the comparison.    
     
     
       42. The method of claim 41, additionally comprising filtering the speech sequence.  
     
     
       43. The method of claim 42, wherein the filtering comprises separating one or more notes of the speech sequence having a duration shorter than a particular duration threshold value.  
     
     
       44. The method of claim 42, wherein the filtering comprises retaining one or more notes of the speech sequence having a loudness above an amplitude threshold value.  
     
     
       45. The method of claim 41, additionally comprising establishing a pitch mean value of a select portion of the speech sequence and using the mean value to identify the plurality of notes of the speech sequence.  
     
     
       46. The method of claim 41, wherein the at least two of the plurality of notes comprise two successive fundamental notes.  
     
     
       47. The method of claim 46, further comprising identifying a frequency change between the two successive fundamental notes.  
     
     
       48. The method of claim 41, wherein the second frequency is calculated from one or more second speech sequences of the individual.  
     
     
       49. The method of claim 41, wherein the second frequency is calculated from one or more second speech sequences of one or more other individuals.  
     
     
       50. The method of claim 41, wherein the second frequency is calculated from the speech sequence of the individual.  
     
     
       51. The method of claim 41, wherein the second frequency further comprises a preferred frequency, and wherein the method additionally comprises presenting the comparison to the individual for speech training of the individual.  
     
     
       52. The method of claim 41, wherein the identifying frequency intervals comprises identifying frequency intervals between at least three fundamental notes in succession.  
     
     
       53. A system, comprising:
 a processor configured to
 identify a plurality of notes of a speech sequence of an individual, 
 identify a plurality of frequency intervals, each frequency interval occurring between a pair of the plurality of notes in the speech sequence, and 
 calculate a frequency of the plurality of frequency intervals in the speech sequence; and 
   a memory configured to store a second frequency, wherein the processor is further configured to compare the calculated frequency with the second frequency and to output one or more results based at least on the comparison.    
     
     
       54. The system of claim 53, further comprising a filter configured to separate notes from the speech sequence that do not exceed at least one of a threshold duration and a threshold amplitude.  
     
     
       55. The system of claim 53, wherein the calculated frequency comprises a proportion of the plurality of frequency intervals to a total number of intervals in the speech sequence.  
     
     
       56. The system of claim 53, wherein the processor is further configured to determine a characteristic of the individual based at least on the comparison.  
     
     
       57. The system of claim 53, wherein the processor is further configured to detect a position of at least one of the plurality of frequency intervals with respect to a second frequency interval.