PocketSphinx is not a program, it is a speech recognition library written in C. It has language bindings for Python, which means you can use the library from Python programs too; to Python programs, the library looks like a normal Python library.

You don't really need to know C to use PocketSphinx, if you know Python. The example program in the tutorial is written in C, but there is an example on how to use PocketSphinx as a GStreamer component with the user interface written in Python.

I don't think you have to use GStreamer to use PocketSphinx from Python programs, it just makes things easier. Without GStreamer, you'd probably have to use ALSA or some audio library to read the raw audio samples, and feed them to the library yourself. Nothing too complicated, but I'd still recommend checking out the PocketSphinx-GStreamer-Python stuff first.

Thanks. I had perused the Gstreamer/Python tutorial a while ago when I knew zero. Now I have a better understanding as to what is going on.

I have dreams of writing a full ISR program using the Sphinx4 recognizer (written in Java) in a robotics application with the goal of passing the Total Turing Test.

Until then, it appears I should pursue the easier route with pocketsphinx and the above tutorial to get some practical programming experience. But time is limited so I would prefer to start with what I will be ending up with. In other words, countless hours would be spent with Gstreamer and Python when the time could be spent learning Java. According to the tutorials, Sphinx4 is a better candidate for the high goal mentioned above. For example, configuration files can be modified on the fly, .i.e, while it is running.

KDE's command and control project died a few years ago along with a few others. It seems distro and hardware compatibility are major issues, not to mention if tts is to be integrated as well. Moreover, audio in Linux is neither easy nor fun. This is frustrating and encouraged me to take on the task.

I thank for your input. You have been very helpful.