SYSTEM AND METHOD FOR MEASURING SPEECH CHARACTERISTICS

• US 20110191104A1
• Filed: 01/29/2010
• Published: 08/04/2011
• Est. Priority Date: 01/29/2010
• Status: Active Grant

• Current Assignee: Rosetta Stone Limited (Rosetta Stone Incorporated)
• Original Assignee: Rosetta Stone Limited (Rosetta Stone Incorporated)
• Inventors: Hacioglu, Kadri, Stanley, Theban, Tepperman, Joseph
• Granted Patent: US 8,768,697 B2
• Time in Patent Office: Days
• US Class Current: 704/236
• CPC Class Codes: G09B 19/04 Speaking with audible presentation of the material to be studied G09B5/04
• CPC Class Codes: G09B 19/06 Foreign languages with audible presentation of material to be studied G09B5/04
• CPC Class Codes: G10L 15/00 Speech recognition G10L17/00 takes precedence

Abstract

A method for measuring a disparity between two speech samples is disclosed that may include determining upon a speech granularity level at which to compare the rhythm of a student speech sample and a reference speech sample; determining a duration disparity between a first speech unit and a second, non-adjacent speech unit in the student speech sample; determining a duration disparity between a first speech unit and a second, non-adjacent speech unit in the reference speech sample; and calculating the difference between the student speech-unit duration disparity and the reference speech-unit disparity.

Claims

1. A method for measuring a disparity between two speech samples, comprising:

recording a reference speech sample spoken by a native speaker;

recording a speech sample spoken by a language student;

generating digital data files representative of the reference and student speech samples, thereby providing a reference speech data file, and a student speech data file, the speech sample data files including information indicative of sound intensity levels as a function of time over the chronological durations of the respective speech samples;

segmenting the reference and student speech data files into respective reference and student speech-unit data files, with each said speech-unit data file including a sequence of speech units over its speech sample duration;

determining a duration disparity between a first speech unit and a second speech unit, that is not adjacent to said first speech unit, within the reference speech-unit data file, thereby providing a reference speech-unit duration disparity;

determining a duration disparity between first and second speech units in the student speech-unit data file that correspond to the locations in the sequence of speech units of the first and second speech units in the reference speech-unit data file, thereby generating a student speech-unit duration disparity; and

calculating a difference between the student speech-unit duration disparity and the reference speech-unit duration disparity.

2. The method of claim 1 further comprising:

scaling at least one of the speech samples so that the total duration of the reference speech sample and the student speech sample are the same.

3. The method of claim 1 wherein the type of speech unit the duration of which is determined is selected from the group consisting of:

phonemes, syllables, and words.

4. The method of claim 1 further comprising:

determining the most granular level of speech in which the number of speech units is the same for the reference speech sample and the student speech sample; and

comparing characteristics of speech units at corresponding locations along the sequences of speech units of the reference and learner speech samples.

5. The method of claim 1 further comprising:

calculating the difference in duration between each speech unit in a sequence of speech units in the student speech sample and each other speech unit in the student speech sample, thereby generating a sequence of student speech-unit duration differences.

6. The method of claim 1 wherein the calculating step comprises calculating differences in the durations of phonemes in the student speech sample.

7. The method of claim 5 further comprising:

calculating the difference in duration between each speech unit in a sequence of speech units in the reference speech sample and each other speech unit in the learner speech sample, thereby generating a sequence of reference speech-unit duration differences corresponding in length to the sequence of student speech-unit duration differences.

8. The method of claim 7 further comprising:

summing the absolute values of the differences between the student speech-unit duration differences and respective reference speech-unit duration differences, to provide a difference sum;

summing the absolute values of the sums of the student speech unit duration differences and respective reference speech unit duration differences, to provide a total absolute difference; and

dividing the difference sum by the total absolute sum to provide the Pairwise variability error (PVE).