Personalized. Friendly. Quality Training.

This site contains three software products that provide a numerical and visual assessment of the prosodic characteristics of speech intonation, namely:

• The analyzer of the statistical characteristics of speech melody - Speech Melody Meter (year of creation 2022);
• The analyzer of the statistical characteristics of the tempo of speech - Speech Rate Meter+ (year of creation 2021);
• The analyzer and trainer of the multi-lingual speech melodic "IntonTrainer" (year of creation 2019);

Users:

The geography of users access to the materials of this site, from the beginning of 2019 to the present, see the site.

Authors:

• Prof. Boris Lobanov - Scientific supervisor
  
  • [email protected]
• Vladimir Zhitko - Software developer
  
  • [email protected]
  • GitHub
  • LinkedIn

Functionality

The Speech Melody Meter (SMM) software is designed to evaluate the statistical characteristics of speech melody by constructing and displaying histograms of the frequency distribution of occurrence of discrete values of the pitch frequency (F0) over sufficiently long speech intervals.

SMM implements a method for estimating three important statistical characteristics of the histogram: the Center of «mass» of the distribution - C [Hz], the Band width - B [Hz], and the Asymmetry distribution - A [Hz].

SMM has a wide range of options for choosing linear scales for constructing histograms used in the analysis of speech melody characteristics, namely: the range of analyzed frequencies - F0min and F0max, as well as frequency resolution - dF0. The choice of specific values of these parameters determines the number of histogram components - N.

SMM is implemented as a separate application for Windows (7, 8, 10).

A distinctive feature of the software implementation of SMM is that the discrete values F0 of speech are measured only at intervals of vowels, which makes it possible to identify the most significant parts of the F0 trajectory for perception and at the same time avoid a number of artifacts that occur at intervals of consonants and pauses.

Areas of use

The use of SMM can be useful in the following practically important cases:

• In scientific research of the different expressive function of speech melodic: expressing the personal accent, normativity, style and artistry;
• When studying and objectively assessing the degree of emotionality of one's speech or the speech of the interlocutor;
• For individual training in preparing speeches and presentations for self-control and teaching the desired characteristics of speech melody;
• As a means of monitoring speech intonation in the real work conditions of call center operators, radio and television announcers.
• As a means of self-control to eliminate speech defects associated with various physiological disorders and diseases;

In the current version of SRM+ from 2021-08-02, unlike the previous SRM from 2021-04-05, a new important function is added to detect "FILLER" sounds. Its use makes it possible to assess the degree of "pollution" of the analyzed speech by sounds of this kind.

Functionality

The "SpeechRateMeter+" software is designed to assess the complex of prosodic characteristics of the speech tempo, namely:

1. Speech rate – Rs – total speed of reading the text (calculated as the number of words per minute during the actual pronunciation time, i.e. taking into account the duration of interphrase pauses);
2. Articulation rate – Ra – phonemic stream generation rate (calculated as the number of words per minute for the total time of articulation, i.e. without taking into account the duration of interphrase pauses);
3. Pauses score – Tp – the mean duration of the inter-phrase pauses (calculated as the RMS value of the duration of all pause sections);
4. Filler score – Tf – the relative duration of "filler" vocal insertions (calculated as a ratio in pae cents of the RMS to the average-median duration of vowels).

Areas of use

SRM+ is implemented as a separate application for Windows (7, 8, 10). The use of SRM+ can be useful in the following practical cases:

• When preparing speeches and presentations for self-control and self-study of the desired rate of speech;
• To improve the skills of oral reading in the native and foreign languages;
• As a means of self-control to eliminate speech defects associated with various physiological disorders and diseases;
• For individual training of oral and speech skills in a number of professions such as: call center operators, radio, TV announcers, etc;
• As a means of monitoring the rate of speech during the work of call-center operators, radio and television announcers in real conditions.

Functionality

IntonTraner is a software package for Windows and Linux, designed for training intonation of oral speech. • Training is based on comparing the melodic portraits of the pattern and pronouncing the phrase to the students. • The acoustic database includes a large set of template phrases covering most of the intonation patterns of the language. • The system is based on interactive auditory and visual contact with the student, and also provides a quantitative assessment of the fidelity of intonation. • Currently, a system has been prepared some basic exercises for training intonation of Russian, British English, as well as basic exercises to improve singing skills. • More over there are several demos for Belarusian, American English, Chinese and German, as well as the intonation of emotional speech. • Localization into other languages ​​can be easily accomplished by creating appropriate acoustic databases.

Areas of use

The software package is recommended for use:

• In linguistic education used as a means of visualizing intonation;
• In individual intonation training for correct pronunciation used as a means of feedback;
• In scientific and practical research used as a means of comparing intonation from different sources.