Hi !

i wanna create speech to text (again), with manual labelling and not using flat-start.

My speech files was labeling manual using waveform and devide per phoneme..

how about flat-start ? what tools i can use to make it?
do u have a sample about that ?

thanks