Характеристика сегментных и просодических единиц в разных типах речи:
норма и реализация


The labelling of the selected phrases (those from spontaneous speech, read text and produced in isolation) was performed by Pavel Skrelin on the basis of segmentation rules used for concatenation of allophones in the allophone-based speech synthesis system. The rules are described in P.Skrelin "Segmentation and Transcription", SPb, 1999 (in Russian).

The first step of labelling was performed with the use of the Wave Assistant: the boundaries between sound segments were marked. The results were converted into TextGrids for Praat. Next steps (transcription) were performed with the use of PRAAT. For the transliteration we used a freeware program found in the Internet (TransLit v. 1.1); this allows us a uniform representation of the orthographic sequence.

TextGrids have the following tiers:

We use the IPA chart, since, as our exsperience has shown, the SAMPA presentation is much more difficult to interprete because of the numerous details we are trying to show.

Symbols used in file names are as follows:

Special symbols and diacritics used in transcription tiers:


  1. 9 - central close-mid unrounded
  2. °3 - central open-mid rounded


  1. ~ - velarized consonant except for velarized l as there is a separate symbol for it in fonts supported by Praat; e. g. ~
  2. () - upper score; e. g. t()
  3. ^ extra-short except for i, a, e, o, u; e. g. æ^
  4. + - voiced consonant; e. g. t+
  5. - - unvoiced consonant as well as the standard symbol; e. g. r-
  6. * - syllabic sonant, e. g. l*