markli - 2011-10-21

Hi, when I am using the sphinx3_conf, I choose the -confoutputfmt s3

sphinx3_conf -confoutputfmt s3

input hypseg:

wav1 S 3146895 T -3045100 A -2938419 L -106681 0 192277 -7675 ++NOISE++ 37
-904371 -11227 I 72 -580925 -9755 CAN 97 -235302 -5859 HELP 128 -103395 -1209
YOU 139 -36752 -4450 WITH 153 -42993 -8829 ME 169 -186089 -7675 <sil> 182
29582 -23633 MAY 199 -311078 -1115 I 211 -814630 -1951 HAVE 231 -251622 -8074
PLEASE 246 -202154 -4177 YOUR(2) 260 -58491 -3377 NAME 295 567524 -7675
++LIPSMACK++ 336 </sil>

output hypseg:
wav1 S 0 T -2945816 A -2938419 L -7397 295 567524 -540 ++LIPSMACK++ 260 -58491
-182 NAME 246 -202154 -249 YOUR 231 -251622 -573 PLEASE 211 -814630 -63 HAVE
199 -311078 6 I 182 29582 -1870 MAY 169 -186089 -540 <sil> 153 -42993 -636 ME
139 -36752 -271 WITH 128 -103395 -1 YOU 97 -235302 -389 HELP 72 -580925 -713
CAN 37 -904371 -836 I 0 192277 -540 ++NOISE++ 0 </sil>

questions:
1) In output, S become 0 while A remains the same. Is it possilbe?
2) the output sequence become reverse, that doesn't matter, but you notice
that the beginning two words both start with frame 0, and the words in whole
sequence start from the frame# of its preceding words should start.

I wonder if this indicating underlying bugs.