labeling system

The uses of Cat_ToBI

of the system

Prosodic phrasing


Cat_ToBI Labeling System

A Cat_ToBI transcription for an utterance consists of an audio recording of the utterance and three windows (see Figure 1 for an example): the waveform of the utterance, the fundamental frequency (F0) contour of the utterance, and the five Cat_ToBI annotation tiers:


1. The Orthographic Tier

This tier contains the orthographic transcription of the text and is equivalent to the Orthographic Tier in the American English ToBI.

2. The Phonetic Transcription Tier

This tier contains the phonetic transcription of the text into the International Phonetic Alphabet (IPA) and is equivalent to the Orthographic Tier in the American English ToBI.

3. The Break Index Tier

The break index tier is based on other ToBI systems, such as MAE_ToBI, Gr-ToBI, etc. There are five break indices, 0, 1, 2, 3 and 4.

  • BI 0 mark cohesion between orthographic words. Orthographic words separated by BI 0 constitute a prosodic word (PrWord) that may bear only one pitch accent.

  • BI 1 mark boundaries between prosodic words (PrWords). Items separated by BI 1 should carry at most one pitch accent each.

  • BI 2 mark either a perceived disjuncture with no intonation effect, or an apparent intonational boundary but with no slowing or other break cues.

  • BI 3 mark the boundaries of ips (intermediate phrases).

  • BI 4 mark the boundaries of IPs (Intonational Phrases).

4. The Tone Tier

For the intonational analysis of Catalan utterances we recognize two types of tonal events, pitch accents and boundary tones, and two levels of phrasing, the intermediate phrase (ip) and the intonational phrase (IP). The following subsections describe the inventory of those tonal events for Catalan, describing their phonetic realizations and their distributional properties. Even though the work presented here is the result of a long experience with the literature, this is to be considered a work in progress and the tonal units described here are tentative. The Cat_ToBI group continues to collect data and to conduct research on several issues related to intonational contrasts in Catalan.

4.1. The pitch accents

Six basic pitch accents have been found in Catalan:

  • 2 monotonal: H* and L*

An explanation about the prototypical phonetic realization and distribution of these accents is described in each one of the links.

4.2. The boundary tones

Eight final boundary tones have been found in Catalan:

and one initial boundary tone: %H

An explanation about the prototypical phonetic realization and distribution of these accents is described in each one of the links.

5. The Miscellaneous Tier

The Miscellaneous Tier has been used for everything from noting non-speech events to commenting on labelling difficulties. Because much of the notation in this tier has not been standardized, it has been of limited use for drawing conclusions from large labelled corpora.

The miscellaneous tier is in essence a ‘comment’ tier for the optional marking of events of any kind other than the standard words, phonetics, tones, and disjunctures marked on the orthographic tier, the phonetic transcription tier, the tone tier, and the break index tier.

Even if our examples in this webpage do not contain this tier, we encourage to use it to note the following events with the corresponding standard labels:

disfluencies such phenomena as stumbling over a word, or abruptly cutting off a word or phrase in midstream to make a fragment phonetic error
lexical self-corrections of parts of sentences repair
lexical self-corrections of whole sentences fresh start
  hesitation pause
differences in rate of speech increase of the rate of speech fast
decrease of the rate of speech slow
cough   cough
laugh   laugh
noises other noises that parasit the data noise

For more information about the use of the Miscellaneous tier, we invite you to visit the website of the MAE-TOBI Guidelines.

Notice that Brugos et al. (2008) introduces one more labelling tier, the alternatives (alt) tier, which uses standard machine-readable notation to explicitly capture the alternative transcriptions considered for an ambiguous region of an utterance.
Alejna Brugos, Nanette Veilleux, Mara Breen, Stefanie Shattuck-Hufnagel, "The Alternatives (Alt) Tier for ToBI: Advantages of Capturing Prosodic Ambiguity", Speech Prosody 2008, Aix-en-Provence.