PHONETIC LABEL FORMATS


The NICO toolkit supports the following phonetic label file formats:

 
timit  The phonetic label format of the TIMIT database. Points in time are referred to by sample number. 16 kHz sample rate.
htk 
 
The htk phonetic label format. Like timit, but all times are multiplied by 625.
mix  The mix format used at KTH. Times are referred to by sample number. 16 kHz sample frequency. Phonetic, phonematic, and word tiers in the same file. 
Example: 
CT 1  
Labels: J'A:Gg V'IL "]:KkA p: T'I V"AKkS#H]LM IDd'A:Gg 
. 
FR 11255 #J >pm #J >w jag 0.703 sec 
FR 12280 $'A: >pm $'A: 0.767 sec 
FR 13624 $G >pm $G 0.851 sec 

 ...

mix_w  The word tier of the mix format (instead of the phonetic symbols after the # or $ characters, the words after ">w" are used).