Hi there,
I need to optimise the speech recognition accuracy of my program. the input is live and at present it is using WSJ model. At present I am trying to hook the program up to the Hub4 model. yet to no avail.
Also when I looked at the build.xml file in the tests/performance/hub4 there was this in it:
hub4_trigram_live * -->
<!-- * * -->
<!-- ********** -->
<target name="hub4_trigram_live"
description="Live mode unsupervised HUB-4 tests">
<fail message="WARNING: This test is not working yet!"/>
<java classpath="${classpath}"
....etc....
The warning message is a concern. Does that mean that hub4 does not work with live recognition yet?
Do you have other suggestions? Because with WSJ the accuracy is close to 0.
Any help is welcome.
Antonia.
PS i'm running sphinx on Eclipse
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi there,
I need to optimise the speech recognition accuracy of my program. the input is live and at present it is using WSJ model. At present I am trying to hook the program up to the Hub4 model. yet to no avail.
Also when I looked at the build.xml file in the tests/performance/hub4 there was this in it:
hub4_trigram_live * -->
<!-- * * -->
<!-- ********** -->
<target name="hub4_trigram_live"
description="Live mode unsupervised HUB-4 tests">
<fail message="WARNING: This test is not working yet!"/>
<java classpath="${classpath}"
....etc....
The warning message is a concern. Does that mean that hub4 does not work with live recognition yet?
Do you have other suggestions? Because with WSJ the accuracy is close to 0.
Any help is welcome.
Antonia.
PS i'm running sphinx on Eclipse