Speech recognition by continuously reading the signal from the microphone?

Question

Sphinx

Comrades, please advise working examples of speech recognition, preferably for python, c ++ or Qt, you can even java.
The task is as follows: the signal from the microphone is continuously read, and when a command is detected, some actions are performed.
I tried the following tools: julius, sphinx, voxforge - the result is not very good. I put Russian morphs in sphinx - it works for a very long time and is very incorrect ...
I don’t want to fence my bikes, so I ask you for advice. A working example is preferred. For example, "julius-3.5.2-quickstart-windows" found on their site works great and fast with the phrases "CALL | DIAL ""0..9" "JOHN...", but it didn't work out to add my phrases there - he wrote that morphemes were not found, the Russians refused to eat, but he worked with a bang - he instantly analyzed the speech and gave the result.
It is desirable working examples - I will be very grateful.
Thanks in advance.

Answer 1

www.nuance.com/ here these guys give sdk they were recognized in siri and google android of old versions.
sphinx is generally an experimental academic library, it will be of little use not for English, because it takes a long time to learn ...