k-tobi labeling conventions

K-ToBI Labeling Conventions

Sun-Ah Jun, Linguistics, UCLA Version 3.1, November 2000http://www.linguistics.ucla.edu/people/jun/ktobi/K-tobi.html Presented by Kyuchul Yoon (Division of English, Kyungnam University) The 2007 Winter Workshop of the Circle of Experimental Phonetics

Background

Korean Tones and Break IndicesA prosodic transcription convention for standard (Seoul) KoreanBased on English ToBI and Japanese ToBI systemAssumes intonational phonology (e.g., Pierrehumbert 1980, Beckman & Pierrehumbert 1986, Pierrehumbert & Beckman 1988)Intonational analysis and prosodic model of Seoul Korean based on Jun (1990, 1993, 1996, 1998; also see Lee 1989 and de Jong 1989 for earlier studies). A first version of K-ToBI was developed in late 1994 by Mary Beckman and Sun-Ah Jun

Intonational Structure (Seoul)

Intonational Structure (Seoul)

Two intonationally defined prosodic units: Intonation Phrase (IP) and Accentual Phrase (AP)IP : marked by a boundary tone (%) (& final lengthening). AP : marked by a phrasal tone (THLH), but not by final lengtheningThe boundary tone is realized in the IP final syllable. At least nine boundary tones identified (L%, H%, LH%, HL%, LHL%, HLH%, HLHL%, LHLH%, LHLHL%)

Structure of K-ToBI

English ToBI has four parallel tiers (1) word, (2) tone, (3) break-index, and (4) miscellaneous, but allows the free proliferation of site-specific extra tiers, e.g. a phones tier for phonetic segmentationCurrent version of K-ToBI expands a tone tier into two : A phonological tone tier: underlying tones A phonetic tone tier : surface tonal patterns

Structure of K-ToBI

Minimally, a K-ToBI transcription for an utterance consists of, (A) a recording of the speech, an associated record of (B) the fundamental frequency contour, and the following five parallel tiers: (1) word (2) phonological tone (3) phonetic tone (4) break-index (5) miscellaneous

Tiers

WordPhonological tonePhonetic toneAP final tonesAP initial tonesIP final boundary tonesBreak indexMiscellaneous

Word tier

Site-specific needsWords may be labeled using either Hangul or romanizationWord as a sequence of segments divided by a space in a written Hangul textThe word label should be placed at the end of the final segment in the word (waveform or spectrogram). That is, each word should be marked at its right edgeFilled pauses, etc. should also be labeled using some site specific convention for the Hangul or romanized spelling

Phonological tone tier

Mark (1) the boundary tone of an IP and (2) the boundary tone of an IP-medial APSince the AP boundary tone in an IP-final position is overridden by an IP-final boundary tone, only IP final boundary tone (%) will be labeled at the end of an IPTo mark the end of an AP, Use LHa as a short term for LHLHa or HHLHaTo mark the end of an IP, Use one of the nine different boundary tones

Phonological tone tier

LHa marks the end of an IP-medial AP, aligned with the end of AP final segment. The LHa tone should be placed at or just before the corresponding break index marker regardless of the actual location of the peak.T% marks the end of an IP, aligned with the end of IP final segment. T can be H, L, HL, LH, HLH, LHL, HLHL, LHLH or LHLHL. A T% tone should be placed at or just before the corresponding break index marker regardless of the actual location of the peak.

Phonetic tone tier

Mark the surface realization of AP tones and IP tonesAP tones: 3 initial tones (i.e. L and H on the 1st syllable, and +H on the 2nd, sometimes the 3rd when the AP is long and focused) 3 final tones (i.e. La and Ha on the final syllable, and L+ on the penult of an AP) When an AP has 3 syllables, the tone on the 2nd syllable can be either L (ex. LLH) or H (ex. LHH).LLH: parsed as L-LH with the undershoot of the 1st H of LHLH, i.e. L L+ Ha LHH: parsed as LH-H with the undershoot of the 2nd L of LHLH, i.e. L +H Ha

Phonetic tone tier

Phonetic tone tier

AP final tones: Ha The most common AP final tone of an IP-medial AP. Either the end of a rising tone or a high flat tone. Aligned with an actual f0 peak on the AP final syllable. La A less common AP final tone, sometimes seen when the following AP begins with a H tone. Aligned with an actual f0 valley on the AP final syllable. L+ Low toned penultimate syllable of an AP. Do not label this tone if predictable, e.g. H (L+) La, L (L+) La. Aligned with an actual f0 valley on the penult of an AP. When only a low plateau, place it at the beginning of the low plateau when preceded by an initial H, or at the end of the plateau when followed by a final H.

Phonetic tone tier

AP initial tones: L On the first syllable of an AP. Aligned with the f0 valley on the first syllable of an AP. H On the first syllable of an AP. Aligned with the f0 peak on the first syllable of an AP. (avoid the first pitch point at the vowel beginning which is most likely due to the segmental perturbation) +H On the second syllable of an AP. (or sometimes the third syllable when the AP is long or uttered fast or produced under focus) Aligned with the f0 peak on the second syllable. When the peak continues over the following syllable, place it aligned with the latest f0 peak of the phrase initial peak.

Phonetic tone tier

Vertical line represents the beginning of the IP-final syllable.

Phonetic tone tier

Boundary tone placement: f0 maximum: H%, LH%, HLH%, LHLH% f0 minimum: L%, HL%, LHL%, HLHL%, LHLHL%Complex boundary tones: HL%, HLH%, LHLH%, LHLHL% Put > at the f0 peak of the non-final H tone.IP type: Determined by the f0 shape of the IP-final syllable * H tone of HL% is sometimes realized on the penultimate syllable of an IP: Old/Middle Korean of dramas, movies, etc.

Phonetic tone tier

IP final boundary tones: L% A level ending, or a gently falling tone over much of the IP-final AP. The most common tone in stating facts, and in declaratives in reading. Placed phrase-finally, aligned with f0 minimum. H% A rising tone beginning to rise before the IP-final syllable, and reaches its peak during the final syllable. The most common tone in seeking information as in yes/no questions. The rise is earlier than that in LH%. Placed phrase-finally, aligned with f0 maximum. LH% A rising tone that is more localized than H%, rising sharply well within the final syllable. Common for questions, continuation rises, and explanatory ending. Placed phrase-finally, aligned with f0 maximum.

Phonetic tone tier

IP final boundary tones: HL% A falling tone that rises to a peak before the last syllable, and then falls during the last syllable. Common in declaratives and wh-questions. Common in news broadcasting. Placed phrase-finally, aligned with f0 minimum. H marked by > aligned with the f0 peak. LHL%, HLH%, LHLH%, HLHL%, LHLHL%

Break index tier

The degree of juncture perceived between each pair of words and between the final word and the silence at the end of the utterance.Marked after all words that have been transcribed in the word tier.All junctures -- including those after fragments and filled pauses -- must be assigned an explicit break index value.

Break index tier

0 For cases of clear phonetic marks of clitic groups; e.g. application of vowel coalescence rules. Also for cases of incomplete nouns; e.g. 1 For phrase-internal word boundaries which are not marked by such cliticization phenomena and can be pronounced by itself. 2 For cases of a minimal phrasal disjuncture, with no strong subjective sense of pause -- that is, a sense of phrase edge of the type that is typically associated with the tonal pattern at the right edge of the AP. 3 For cases of a strong phrasal disjuncture, with a strong subjective sense of pause (whether it be an objective visible pause or only the virtual pause cued by final lengthening) -- that is, a sense of phrase break of the type that is typically associated with the tonal pattern at the right edge of an IP.

Break index tier

In case of mismatch btw/ tonal markings and break indices, the break index number should follow the perceived juncture rather than the tones. 1m A disjuncture that typically would correspond to a phrase-medial word boundary, but is marked by the tonal pattern of an AP. 2m A medium strength disjuncture that typically would be marked by the tonal pattern of an AP, but without any tonal markings, or with those of an IP. 3m A highest strength disjuncture that typically would be marked by the tonal pattern of an IP, but with the tonal markings of an AP. # - Break uncertain between # and #-1 level #p Pause or disfluency after this level of juncture

Miscellaneous tier

For any comments or markings, aligned with both their temporal beginnings and ends. , , ,

Romanization convention

?

?

?

?

() .

sound-loader.praat############################################################# File randomizer for gating experiments # Written by Kyuchul Yoon ([email protected])# The script reads in all the files from the current directory, # sorts them and load them in Praat Object Window############################################################

Create Strings as file list... fileList *.wavSortnumFiles = Get number of strings#pause 'numFiles' files identified. Continue?

# Load the filesfor i to numFilesselect Strings fileListstrFileName$ = Get string... iRead from file... 'subFolderToProcess$'\'strFileName$'endfor

############ END OF SCRIPT ###############

ExtraFilesFromProfSunAhJun/Labelling data.docDrama1 yuNgizaniMnazERdAhwasiRgoQgAaNhanINgEDaRgoiDzyo

L+H L%L LaL HaH HaL HaL+HL+HaL+HL+HL%

32222022-3

Drama2aPahaNtesAQgiNnuguNgarIRnAgasirEhaniKaaPadonahaNteKoyEDzanayo

L+H HaH LaL+HL HaH+H L+HL%L+HL+HaH+H L%

22123-123

Drama3 zEgi,miaNhaNde,azESigaiTagazENhwarIRdasihA zuRKe.

L LH%L+H L+HL%L+H HaL+H LaL+H HaL HaH+H LH%

33-22p2-23

Drama4 YAhagopyEQsAQIRsaRRyEmYnINnizazoNsimINgERegadweyahalgEda.

L+H HaH+H LaH L+HL%L HaL L+HaL+HL%

22-323-13

Interv1gI, nozohagosimiNdaNcehagodoQiRsENsaQenokomaRhanIN gENzoMmuNzegaiSIMnida.

L LaL+H L+HaH+H L+HL%L+H

L+HaHL+Ha L+H L-HL%

2p2311212-3

Interv2zENguGesEzENhwarIRzoMzusiBsiyo.goQsamipalsamirecENbAGsibiRbENiMnida.

L L+HaL+H

L+HL%L L+HaH L+H%H LaH+HL+H%

212-3-23203

Interv3SiregizoQryaQzegamuEsiNzisERmyEQbytaGdIriRKeyo.

H+H HaL+H LaL+H L+LaH+HL+H%

223-2-3

Interv4hyoyuRzEgigohaMnizEgiNgEsIRmaNdIrEzigehanINdeyo.

H+H L+HaH+H L+HaL+HL+HaH+H L+HL%

2022-23

Reading1daSAmanetuhaQhadIDnakyaQhaRsubaKeEBSEDTENgIege

L+H L+HaH+H L+HL%L+H

L+HaL HaL HL%

2301223

Reading2naMsEQUnaMsEQzibAUiRbuiRcEzegasizaGdweES-IMnida.

L+H HaL+H L+HaL HaL+H L+H%H HaL+H L%

2223-1m3

Reading3gatINcEziUdoQnyoyAgiwahaMKewENmAciNhanIroiRgwaNhAwaSESTa.

L Ha H HaL+HHaL H%L+H HaL L+H%LL%

2-212-3232m3

(Reading3 file is very ambiguous)

Reading4dAbubuNUguGKaUhENbEbinagazoGPEBTIQIRbomyEN,

H HaL+H LaH+H L%L+HHaL L%

223123

(In Reading4 file, the first word and the penultimate word are ambiguous)

Story1baraMgwahANnimisErohimidE sedagodatugoiSIRTA,

L HaH L+HL%L HaH HaL La H L+HaL+H L+HL%

23222- 213

Story2gIdIrINnugudINzinagIneUweturIRmENzEbEDginINiga

L HaL+H HaL+H HaL HaL HaL+HL+HL%

2222213

Story3bupuQINhiMKEDburESInabuRmyENbuRSuroGnagInenINweturIRdaNdaNhiyEmyESIMnida

L+H HaH HaL L+HL%L HaL HaL+H L+HL%L HaL L+HaL+H L%

2232-23223

Story4iTEhaNnimiTigEwuNhADbicIRgamaNhinAriCweni,

L HaH L+HaH+H LH%H HaL L+HaL+H LH%

223223

News1guGbaQbunINyEguNzaQgyowahasakwaNinryEgIRkIgenIRrinINdIQ,

L+H L+HL%L HaL HaH+H L+HaH HaL+H LH%

33-212213

News2idIrihyENzAaNzENhaNziyEgedApihAiSImyE, gENgaQdoyaQhohaNgEsIroboiNdagobaRkyESIMnida.

L+H HaH L+HL%L+HL+HaL+H L%L L+HL%L+H

L+HL%H L%

232-2133-0133

New3siByuGTecoQsENiRcagoQcENzarIRhwaGzEQhAbaRpyohASIMnida.

HHaH L+HL%L HL%H La(L) L+HL%L+H L%

123-3-2-3-3

New4goQmuwEnUiNsagwaNrigaziGgIBzuQsimesEziQmuzuQsimIrobaKwigedwe,

L L+HaL HaL+H L%L HaL+H L%L+HL+HL%L+HL%

22-3232m313

(the boundary in zuQsimesE could be HL% but H is realized earlier than the final syllable. This is not uncommon in News style)

ExtraFilesFromProfSunAhJun/sound-loader.praat############################################################# File randomizer for gating experiments # Written by Kyuchul Yoon ([email protected])# The script reads in all the .wav files from the current directory, randomizes them, and#combines them all into one big file that can be burned into a CD############################################################

# Get the user inputform Select files and parameterscomment Add some more options?word fileToProcess_(only_extensions) wavword subFolderToProcess soundsendform

# Randomize all the wav filesCreate Strings as file list... fileList 'subFolderToProcess$'\*.'fileToProcess$'numFiles = Get number of strings#pause 'numFiles' files identified. Continue?

# Printout the list of randomized files. The order should be the same as that in the final CD filefor i to numFilesselect Strings fileListstrFileName$ = Get string... iRead from file... 'subFolderToProcess$'\'strFileName$'endfor

############ END OF SCRIPT ###############

ExtraFilesFromProfSunAhJun/README.FIRST.txtI am attaching 12 wave files and a Word (doc) file where the labels of each sound file are listed.?I will attach 8 wave files at another email.?These 20 speech files are those used for testing the K-ToBI agreement (the agreement data are published in the?ICSLP proceedings paper in 2000,?authored by?Jun, Lee,?Kim & Lee). Please segment a word boundary and label the tones in a relevant location on the tones tier, and let me know if you have any questions.? Some of the files (words) are quite ambiguous for the phrasing. I made those cases in bold.?

ExtraFilesFromProfSunAhJun/sounds/DRAMA1.WAV




ExtraFilesFromProfSunAhJun/sounds/INTVIEW1.WAV




ExtraFilesFromProfSunAhJun/sounds/NEWS1.WAV




ExtraFilesFromProfSunAhJun/sounds/READ1.WAV




ExtraFilesFromProfSunAhJun/sounds/STORY1.WAV




FlashTest/Image1.jpg

FlashTest/2syllAP-LHa.wav

FlashTest/Movie1.sbk

FlashTest/Movie1.swf

FlashTest/K-ToBI.html

FlashTest/Movie1.swi

FlashTest/Movie2.swi

FlashTest/Movie2.swf

data/T1P2S8-1m.wav

data/4boundary-H%.wav

data/4boundary-HL%.wav

data/4boundary-LH%.wav

data/4boundary-LHL%.wav

data/5syllAP-HHLHa.wav

data/5syllAP-LHLHa.wav

data/6syllAP-LHLHa.wav

data/break-L8C3.wav

data/coQgaG-HLH%.wav

data/gazEQgyosa.wav

data/IPboundary-HL%.wav

data/IPboundary-LH%.wav

data/J3A2-HLH%.wav

data/millennium.wav

data/millennium-final.wav

data/millennium-middle.wav

data/T1P1S2.wav

data/T1P1S2-late.wav

data/T1P2S10.wav

data/T1P2S5.wav

data/T1P2S5-late.wav

data/T1P2S6.wav

data/2syllAP-LHa.wav

k-tobi labeling conventions

Documents

ipfinal boundary tone

ap boundary tone

lha tone

phonetic tone tier

end of ip final segment

ipfinal position

ip final syllable

end of ap final segment