Posts: 44
Joined: Fri Apr 05, 2013 1:20 am

Re: LiveSpeechRecognizer ?

Tue Feb 21, 2017 5:53 am

In linux I think you're outta luck. :( as continuous speech recognition requires continuous processing of audio stream from mic. In windows we've got NAudio library but it won't work in Linux OS.

you'll have to implement a custom algorithm to detects changes in audio stream from mic and start speech recognition.

User avatar
Help & Support
Posts: 320
Joined: Fri Oct 25, 2013 9:20 pm

Re: LiveSpeechRecognizer ?

Thu Feb 23, 2017 5:53 am

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. ;)

Return to “General Discussion”

Who is online

Users browsing this forum: No registered users and 0 guests