Transcription

Transcription is the representation of perceived temporal segments of speech with conventional alphabetic symbols. This definition covers the following main kinds of transcription:

1. Orthographic transcription: representation of segments with standard orthography; very close to the procedure of `dictation', except that additional pseudo-orthographic conventions are sometimes introduced for dealing with unconventional spoken language phenomena such as discourse particles, hesitation markers, or fragmentary speech. Orthographic transcription uses the symbols of standard orthographies.
2. Phonemic transcription: representation of segments with a minimum amount of phonetic information consistent with distinguishin (contrastinv) words by substituting these segments. Phonemic transcription uses the symbols of the International Phonetic Alphabet, encoded for computer use either as special fonts (e.g. TrueType fonts for Windows computers, Metafont fonts for UNIX), or coded with numbers (e.g. the Esling numbers) or with ASCII codes (e.g. the SAMPA alphabet).
3. Phonetic transcription: representation of segments with more than the minimum amount of phonemic information; broad phonetic transcription approaches the minimum phonemic type of transcription, while narrow phonetic transcription may employ additional diacritics to mark subtle differences in pronunciation. Phonetic transcription uses the symbols of the International Phonetic Alphabet, as with phonemic transcription.
4. Prosodic transcription: representation of temporal and fundamental frequency patterns in spoken language, including rhythms and accentuation. There are several systems of prosodic transcription, from the elementary system provided by the International Phonetic Alphabet, to the Tonetic Stress Mark (TSM) system, the four-level system of Pike and other US linguists, the INTSINT system of focal pitch movements, and the ToBI system of Tones and Break Indices.

An overview of the principles of transcription and their application to the creation of resources for the spoken language technologies is given in .