Hello, <br><br>I need HTK lattice in my experiments but the sphinx3
decoder I used, could not generate HTK lattice. So I have to convert
sphinx lattice to HTK lattice.<br><br>My problem is : the lattice
generated by sphinx3 decoder provides only the acoustic score of word
transitions, I did not find the option to obtain the lmscore in sphninx
lattice.<br>
<br>In order to obtain HTK lattice with lmscore, first I converted
sphinx lattice to HTK SLF lattice format (I added l=0 as lmscore, the
acoustic score is kept as it is)<br><br>Then I used lattice tool (Srilm V 1.5.2) to rescore the lattice by giving a LM :<br>
<br>> lattice-tool -in-lattice in.slf -read-htk -lm <a href="http://lm.bo/" target="_blank">LM.BO</a> -htk-lmscale 9.5 -htk-wdpenalty 0.7 -htk-logbase 1.0003 -out-lattice out.slf -write-htk<br><br>(the
lmscale, wdpenalty and logbase are the values that I used during
lattice generation with sphninx3, the LM is the same as in sphinx3)<br>
<br>I obtained in the output a lattice with acoustic score and new
lmscore. What I observed is that the acoustic score in the output
lattice is recalculated using the logbase. <br><br>In order to verify
that the output lattice in HTK format is equivalent to the orginal
sphinx lattice once, I generated 200-Best lists from these two lattices.<br>
<br>- for sphinx lattice I used sphinx3_astar to generate N-best<br>- for the rescore HTK lattice, i used lattice-tool : <br><br>
>lattice-tool -in-lattice out.slf -read-htk -lm lm.BO -htk-lmscale
9.5 -htk-wdpenalty 0.7 -htk-logbase 1.0003 -nbest-decode 200
-out-nbest-dir OUT/<br>
<br>The problem is that the order of the hypothesis in the two N-best list is not
the same. The 1-best given by sphinx3_astar ccould be found
in the 200-Best given by lattice-tool but with a much more lower rank or some time not found. But I always find the 1-best of sphinx_astar in a
bigger N-Best list of lattice-tool (N=2000).<br>
<br>I am convinced that this is a problems of normalizing the score
between sphinx and lattice-tool. If the score is correctly normalized,
I should have the same N-best at both sides.<br><br>Could you please give me any clues on this issue?<br>
<br>Thank in advance.<br><br>Sopheap<br clear="all"><br>-- <br>---------------------------------------------<br>Sopheap SENG<br><br>Laboratoire d'Informatique de Grenoble (LIG)<br>Equipe GETALP Bureau C118 <br>220, avenue de la Chimie<br>
Campus Scientifique, BP53<br>38041 GRENOBLE Cedex 9, FRANCE<br>Tél : (33)-4-76-63-55-81 <br>Télécopie : (33)-4-76-63-55-52<br>Courriel : <a href="mailto:sopheap.seng@imag.f" target="_blank">sopheap.seng@imag.f</a><br>URL : <a href="http://www-geod.imag.fr" target="_blank">http://www-geod.imag.fr</a><br>
---------------------------------------------<br>Enseignant<br>Institut de Technologie du Cambodge<br>BP 86, Bd de Pochentong<br>Phnom Penh - Cambodge<br>Tél : (855)-23-88-03-70/98-24-45<br>Télécopie : (855)-23-88-03-69<br>
Courriel : <a href="mailto:sopheap.seng@itc.edu.kh" target="_blank">sopheap.seng@itc.edu.kh</a><br>URL : <a href="http://www.itc.edu.kh" target="_blank">http://www.itc.edu.kh</a><br>---------------------------------------------