Reference " Phone-level pronunciation scoring and assessment for interactive language learning " by S.M. Witt and S.J. Young
it is mentioned for calculating GOP ,
" Firstly, the numerator of is computed using a forced alignment in which the
sequence of phone models is fixed by the known transcription and secondly, the denominator is determined using an unconstrained phone loop the denominator is determined using an unconstrained phone loop".
kindly explain unconstrained phone loop means
1. decoding using phone recognizer or
2. force-aligning the segment with all the phones in database
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
how to get phone loop output in sphinx3.
whether phone loop recognizer can be done by decoding with phone dictionary , without LM and Lattices for phone loop
whether 'segascr' in force align and acoustic score in decoder is same
if so, but the number of frames won't be same in both then how to compare
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Reference " Phone-level pronunciation scoring and assessment for interactive language learning " by S.M. Witt and S.J. Young
it is mentioned for calculating GOP ,
" Firstly, the numerator of is computed using a forced alignment in which the
sequence of phone models is fixed by the known transcription and secondly, the denominator is determined using an unconstrained phone loop the denominator is determined using an unconstrained phone loop".
kindly explain unconstrained phone loop means
1. decoding using phone recognizer or
2. force-aligning the segment with all the phones in database
Unconstrained phone loop corresponds to decoding using phone recognizer
whether phone loop recognizer can be done by decoding with phone dictionary , without LM and Lattices for phone loop
if so, but the number of frames won't be same in both then how to compare