You've literally spammed the Syn Speech forum
keep it cool mate. I am not the developer of Syn Speech but here's my 2 cents.
The ASR engine is designed to deal with audio files mainly. That is why even the sample project does the same by first saving the mic audio stream to file and then running the recognizer.
I ain't an expert in audio stream recording especially in Linux environment but I myself have used it in android. I used Xamarin's MediaRecorder
class to record speech and send the transcribed text values to my remote application as I didn't want to be bounded by Androids Speech Recognition API limits. Just because you aren't able to implement an audio stream interface you cannot dismiss its usability.
If Syn Speech it just cutting for you, here's a trivia, if you are really into getting an asr interface for your project why not compile pocketsphinx
and read the output stream of the console application at real time. I have done that is the past and it too worked well for me with some minor exceptions. There's also BASS.NET API that you can fiddle around with and see if that helps you in recording audio streams from mic in Linux.
hey n by the way if you find my post helpful why not get rid of the other posts you've made in frustration.