BNF style CAG format definition =============================== CAGFILE := DICFILE | ANNFILE | COMBOFILE DICFILE := DICTIONARY DICTIONARY := A | A nl T | A nl G | A nl T nl G A := "#A" nl AGSETID | "#A" nl AGSETID nl METADATA T := "#T" nl TIMELINEID_LOCAL | "#T" nl TIMELINEID_LOCAL nl S | "#T" nl TIMELINEID_LOCAL nl METADATA | "#T" nl TIMELINEID_LOCAL nl METADATA nl S | T nl T S := "#S" nl SIGNALID_LOCAL URI MIMECLASS MIMETYPE ENCODING UNIT | "#S" nl SIGNALID_LOCAL URI MIMECLASS MIMETYPE ENCODING UNIT TRACK | "#S" nl SIGNALID_LOCAL URI MIMECLASS MIMETYPE ENCODING UNIT nl METADATA | "#S" nl SIGNALID_LOCAL URI MIMECLASS MIMETYPE ENCODING UNIT TRACK nl METADATA | S nl S G := "#G" nl AGID_LOCAL TIMELINEID_LOCAL nl V | "#G" nl AGID_LOCAL TIMELINEID_LOCAL nl METADATA nl V | G nl G V := "#V" nl UNIT TIMELINEID VSUB | V nl V VSUB := ANCHORID_LOCAL | ANCHORID_LOCAL OFFSET | VSUB nl VSUB METADATA := METADATANAME METADATAVALUE | METADATA nl METADATA ANNFILE := I nl ANNOTATION I := "#I" nl DICFILENAME | I nl DICFILENAME ANNOTATION := ML nl MT nl "#E" nl E ML := "#ML" nl FEATURENAME FEATURENAME_INDEX | ML nl FEATURENAME FEATURENAME_INDEX MT := "#MT" nl ANNOTATIONTYPE ANNOTATIONTYPE_INDEX | MT nl ANNOTATIONTYPE ANNOTATIONTYPE_INDEX E := AGID_LOCAL ANNOTATIONID_LOCAL ANCHORID_LOCAL ANCHORID_LOCAL ANNOTATIONTYPE_INDEX ";" F | E nl E F := FEATURENAME_INDEX FEATUREVALUE | F ";" F COMBOFILE := DICTIONARY nl ANNOTATION nl := new line character TIMELINEID_LOCAL := TimelineId minus AGSetId SIGNALID_LOCAL := SignalId minus TimelineId AGID_LOCAL := AGId minus AGSetId ANNOTATIONID_LOCAL := AnnotationId minus AGId ANCHORID_LOCAL := AnchorId minus AGId "DICFILE" sample (d.dic) ======================== #A Timit #T T1 #S 1 LDC93S1/train/dr1/fjsp0/sa1.wav audio wav wav 16kHz #G 1 T1 #V 16kHz Timit:T1:1 1 0 2 3050 3 4559 4 5723 5 6642 6 8772 7 9190 8 10337 9 11517 A 12500 B 12640 C 14714 D 15870 F 16334 G 18088 H 20417 I 21199 J 22560 K 22920 L 23271 M 24229 N 25566 O 27156 P 28064 Q 29660 R 31719 S 33360 U 33754 V 34715 W 36080 X 36326 Y 37556 Z 39561 a 40313 b 42059 c 43479 d 44586 e 46720 f 46797 "ANNFILE" sample (a.ann) ======================== #I d.dic #ML label 0 #MT phn 0 txt 1 wrd 2 #E 1 E1 1 2 0;0 h# 1 Ee 1 4 2;0 she 1 Ep 1 f 1;0 She had your dark suit in greasy wash water all year. 1 E2 2 3 0;0 sh 1 E3 3 4 0;0 ix 1 E4 4 5 0;0 hv 1 Ef 4 8 2;0 had 1 E5 5 6 0;0 eh 1 E6 6 7 0;0 dcl 1 E7 7 8 0;0 jh 1 E8 8 9 0;0 ih 1 Eg 8 9 2;0 your 1 E9 9 A 0;0 dcl 1 Eh 9 F 2;0 dark 1 EA A B 0;0 d 1 EB B C 0;0 ah 1 EC C D 0;0 kcl 1 ED D F 0;0 k 1 EF F G 0;0 s 1 Ei F I 2;0 suit 1 EG G H 0;0 ux 1 EH H I 0;0 q 1 EI I J 0;0 en 1 Ej I J 2;0 in 1 EJ J K 0;0 gcl 1 Ek J P 2;0 greasy 1 EK K L 0;0 g 1 EL L M 0;0 r 1 EM M N 0;0 ix 1 EN N O 0;0 s 1 EO O P 0;0 ix 1 EP P Q 0;0 w 1 El P S 2;0 wash 1 EQ Q R 0;0 ao 1 ER R S 0;0 sh 1 ES S U 0;0 epi 1 Em S Y 2;0 water 1 EU U V 0;0 w 1 EV V W 0;0 ao 1 EW W X 0;0 dx 1 EX X Y 0;0 axr 1 EY Y Z 0;0 ao 1 En Y a 2;0 all 1 EZ Z a 0;0 l 1 Ea a b 0;0 y 1 Eo a d 2;0 year 1 Eb b c 0;0 ih 1 Ec c d 0;0 axr 1 Ed d e 0;0 h# "COMBOFILE" sample (c.cag) ========================== #A TIMIT22 #T T1 #S 1 LDC93S1/train/dr1/fjsp0/sa1.wav audio wav wav 16kHz #G 1 T1 #V 16kHz TIMIT22:T1:1 1 0 2 3050 3 4559 4 5723 5 6642 6 8772 7 9190 8 10337 9 11517 A 12500 B 12640 C 14714 D 15870 F 16334 G 18088 H 20417 I 21199 J 22560 K 22920 L 23271 M 24229 N 25566 O 27156 P 28064 Q 29660 R 31719 S 33360 U 33754 V 34715 W 36080 X 36326 Y 37556 Z 39561 a 40313 b 42059 c 43479 d 44586 e 46720 f 46797 #ML label 0 #MT phn 0 txt 1 wrd 2 #E 1 E1 1 2 0;0 h# 1 Ee 1 4 2;0 she 1 Ep 1 f 1;0 She had your dark suit in greasy wash water all year. 1 E2 2 3 0;0 sh 1 E3 3 4 0;0 ix 1 E4 4 5 0;0 hv 1 Ef 4 8 2;0 had 1 E5 5 6 0;0 eh 1 E6 6 7 0;0 dcl 1 E7 7 8 0;0 jh 1 E8 8 9 0;0 ih 1 Eg 8 9 2;0 your 1 E9 9 A 0;0 dcl 1 Eh 9 F 2;0 dark 1 EA A B 0;0 d 1 EB B C 0;0 ah 1 EC C D 0;0 kcl 1 ED D F 0;0 k 1 EF F G 0;0 s 1 Ei F I 2;0 suit 1 EG G H 0;0 ux 1 EH H I 0;0 q 1 EI I J 0;0 en 1 Ej I J 2;0 in 1 EJ J K 0;0 gcl 1 Ek J P 2;0 greasy 1 EK K L 0;0 g 1 EL L M 0;0 r 1 EM M N 0;0 ix 1 EN N O 0;0 s 1 EO O P 0;0 ix 1 EP P Q 0;0 w 1 El P S 2;0 wash 1 EQ Q R 0;0 ao 1 ER R S 0;0 sh 1 ES S U 0;0 epi 1 Em S Y 2;0 water 1 EU U V 0;0 w 1 EV V W 0;0 ao 1 EW W X 0;0 dx 1 EX X Y 0;0 axr 1 EY Y Z 0;0 ao 1 En Y a 2;0 all 1 EZ Z a 0;0 l 1 Ea a b 0;0 y 1 Eo a d 2;0 year 1 Eb b c 0;0 ih 1 Ec c d 0;0 axr 1 Ed d e 0;0 h#