I found your Dutch acoustic model on this site and wondered if it can be used also in the java environment of Sphinx 4. I tried to convert the Dutch acoustic model to Sphinx 4 following the steps described here http://cmusphinx.sourceforge.net/sphinx4/doc/UsingSphinxTrainModels.html
I'm not sure this works. Has anyone ever tried to use the acoustic model in Sphinx 4 or is it only usable in Sphinx 3?
>I'm not sure this works. Has anyone ever tried to use the acoustic model in Sphinx 4 or is it only usable in Sphinx 3?
I have not tried it. nsh would be the best person to answer your question - he created the acoustic models you are asking about.
Regardless, this page on the Sphinx4 site seems to indicate that you should be able to do it:
Currently, Sphinx-4 uses models created with SphinxTrain, also available at cmusphinx.org. SphinxTrain generates acoustic models in the format used by Sphinx-3. To create a package as used by Sphinx-4, please check the page about using SphinxTrain models.
Can anyone please tell me how i can use Sphinx4 for free speech recognition in the case of large vocabulary? I have been succesfuly runing the demos, but they are just demos. and limited to digits or specific sentences. which file in the Sphinx4 package i will have to run in order to do speech recognition where the user is allowed to say anything.
your help will be highly appreciated,
Please take a look at Evaldictator project:
Also you need to read Programmers Manual and configuration files in sphinx4/performance/hub4. I hope that helps.
Install SphinxTrain. Look at SphinxTrainDirectory/doc/tinydoc.txt
for new speech training.
You can obtain mixture_variance, means, transition_matrices, variance and *.mdef (and *.dic ) file needed by sphinx4 to make a new demo.