I tried to run julian of the quick-start-project. It runs, so you can see <<< please speak >>>. But It doesn't recognice when I am speaking in my microphone (I am using a USB head-phone with microphone). According to the system settings, it recognizes my microphone and it actually works. but not with julian. Do you guys would know what is wrong?
I also tried to do the tutorial for creating an own speaker dependent acoustic model. If I run it after the last step. I get following failure message. But I can't imagine what could be wrong. I do point to the right. julian.jconf file.
include config: ./julian.jconf
./julian.jconf: wrong argument: -dfa
Try `-help' for more information.
I am using Ubuntu 10.04 LTS - the Lucid Lynx
>But It doesn't recognice when I am speaking in my microphone (I am using
>a USB head-phone with microphone)
You might try it with the newest version of Julius from the the Julius web site.
Note that Julius was once used just for dictation using a statistical language model, and Julian was for grammar based language model. Julian was merged with Julius in version 4.0, so there no longer a separate executable called Julian.
>./julian.jconf: wrong argument: -dfa
You need to confirm what you put in your julian.jconf file (in the .dfa parameter) is correct