Blame view
tools/sctk-2.4.10/src/md-eval/test/md_test4.ref.rttm
3.68 KB
8dcb6dfcb
first commit
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 |
;; mdeval md_test4 ref rttm ;; This test file has some odd speaker names, ;; and will test that md-eval counts Nref for ;; "SU exact detection statistics in terms of words" ;; the same as Nref for "SU detection in terms of # of SU's" ;; This is an important test of getting that logic correct. SPKR-INFO md_test4 1 <NA> <NA> <NA> child spkr1r <NA> SEGMENT md_test4 1 10.0 6.0 <NA> <NA> spkr1r <NA> SPEAKER md_test4 1 10.0 6.0 <NA> <NA> spkr1r <NA> SU md_test4 1 10.0 3.0 <NA> statement spkr1r <NA> LEXEME md_test4 1 10.0 1.0 firstWord lex spkr1r <NA> LEXEME md_test4 1 11.0 1.0 T.'s alpha spkr1r <NA> LEXEME md_test4 1 12.0 1.0 thirdWord acronym spkr1r <NA> SU md_test4 1 13.0 3.0 <NA> question spkr1r <NA> LEXEME md_test4 1 13.0 1.0 fourthWord interjection spkr1r <NA> LEXEME md_test4 1 14.0 1.0 fifthWord propernoun spkr1r <NA> LEXEME md_test4 1 15.0 1.0 sixthWord other spkr1r <NA> ;; Can the scanner handle this oddly formatted speakername? SPKR-INFO md_test4 1 <NA> <NA> <NA> unknown name="<NAr>" <NA> SEGMENT md_test4 1 16.0 3.0 <NA> <NA> name="<NAr>" <NA> SPEAKER md_test4 1 16.0 3.0 <NA> <NA> name="<NAr>" <NA> ;; In the sys file, the last word of this SU will be omitted SU md_test4 1 16.0 3.0 <NA> backchannel name="<NAr>" <NA> LEXEME md_test4 1 16.0 1.0 seventhWord lex name="<NAr>" <NA> LEXEME md_test4 1 17.0 1.0 eighthWord lex name="<NAr>" <NA> LEXEME md_test4 1 18.0 1.0 ninthWord lex name="<NAr>" <NA> ;; How about this speakername? SPKR-INFO md_test4 1 <NA> <NA> <NA> unknown spkr=<sp3r> <NA> SEGMENT md_test4 1 19.0 11.0 <NA> <NA> spkr=<sp3r> <NA> SPEAKER md_test4 1 19.0 11.0 <NA> <NA> spkr=<sp3r> <NA> ;; In the UEM file, the last 1.51 words of this SU will be no-eval'd ;; and in the sys file the last two words of it will be omitted SU md_test4 1 19.0 5.00 <NA> question spkr=<sp3r> <NA> LEXEME md_test4 1 19.0 1.0 tenthWord lex spkr=<sp3r> <NA> IP md_test4 1 20.0 <NA> <NA> filler spkr=<sp3r> <NA> FILLER md_test4 1 20.0 1.0 <NA> filled_pause spkr=<sp3r> <NA> LEXEME md_test4 1 20.0 1.0 eleventhWord fp spkr=<sp3r> <NA> LEXEME md_test4 1 21.0 1.0 twelfthWord lex spkr=<sp3r> <NA> LEXEME md_test4 1 22.0 1.0 thirteenthWord lex spkr=<sp3r> <NA> LEXEME md_test4 1 23.0 1.0 fourteenthWord lex spkr=<sp3r> <NA> ;; The following ref SU is a duplicate of the preceding ref SU ;; In the UEM file, the last 1.50 words of this SU will be no-eval'd ;; rather than the last 1.51 ;; and in the sys file the last two words of it will be omitted SU md_test4 1 24.0 5.00 <NA> question spkr=<sp3r> <NA> LEXEME md_test4 1 24.0 1.0 fifteenthWord lex spkr=<sp3r> <NA> IP md_test4 1 25.0 <NA> <NA> filler spkr=<sp3r> <NA> FILLER md_test4 1 25.0 1.0 <NA> filled_pause spkr=<sp3r> <NA> LEXEME md_test4 1 25.0 1.0 sixteenthWord fp spkr=<sp3r> <NA> LEXEME md_test4 1 26.0 1.0 seventeenthWord lex spkr=<sp3r> <NA> LEXEME md_test4 1 27.0 1.0 eighteenthWord lex spkr=<sp3r> <NA> LEXEME md_test4 1 28.0 1.0 nineteenthWord lex spkr=<sp3r> <NA> SU md_test4 1 29.0 1.0 <NA> question spkr=<sp3r> <NA> LEXEME md_test4 1 29.0 1.0 twentiethWord lex spkr=<sp3r> <NA> |