[Freeswitch-users] Speech to text voicemail

Hi, I want to use Sphinx for voicemail transcription. I have successfully setup the Sphinx using standard language module hub4.5000.DMP, acoustic module hub4wsj_sc_8k and dictionary cmu07a.dic. I tried this setup but the accuracy was very less. I have read that we need to use custom language module for improving the accurary (as the default one is the universal). It will be great if anyone had worked on this and help me out with this. -- Thanks and Regards, Ashwin Jain