Questions[speech-recognition](ubuntu)

I'm looking for software that allows live dictation into a word processor like WPS Writer. It would also be great if it could somehow distinguish between background talking - e.g. a busy kitchen and my voice.

I don't have a specialised microphone, except for some cheap earplugs and the built-in Dell microphone.

I'd like it to be either offline completely or not a web app.

Are there any programs that will let me dictate (speech-to-text) to any program? Or at least to a text editor and text fields in Firefox.

I’m not interested in anything that uses an online service, it has to be on-device. It may be a commercial product.

I have a bunch of audio files contained ordinary speech, mainly in mp3 format which I'd like to convert to text. Does anyone know of software that can do this?

Here is what i want to do:

1. I want to put/run Orca as an idle process in background after booting into Ubuntu. (No problem there so far.)

2. Now when I open up a text file or website in my [web]browser; I want to be able, to highlight text and use a keyboard-shortcut in order to invoke Orca process [see step 1], to read the selected text back to me.

My problem is, that Orca, once initialized as a running process permanently reads back everything, from mouse-over actions of buttons, to system related actions etc.. and orca-preferences doesn't come with options, to reduce and suppress this. Orca might not be the right application at all for this and a common speech-synthesis application might be even better.

• If someone knows how to make it work that way, please help.
• An alternative speech-synthesis "read back" option/function, like Apple offers in OS X would even be better, since there one can simply "selected text", that is invoked with Ctrl+# shortcut, or simply use say, or for text files say -f ~/input.txt -o ~/output.aiff-command to have text red back instantly!

Is there a Ubuntu alternative for this program?