LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Linux - Software (https://www.linuxquestions.org/questions/linux-software-2/)
-   -   Speech Recognition Software for Linux (https://www.linuxquestions.org/questions/linux-software-2/speech-recognition-software-for-linux-4175554956/)

Ztcoracat 09-30-2015 08:58 PM

Speech Recognition Software for Linux
 
I've been doing some research on Speech Recognition Software and I'm having trouble finding any concrete evidence of folks saying that it works efficiently and has good functionality.

I heard about CMU Sphinx a few years ago.
http://cmusphinx.sourceforge.net/

My Python buddy told me about Simon but he mentioned he would like to see Simon become a reality.
https://wiki.archlinux.org/index.php...h_recognition\


Like many of you if you type a lot and have been typing for many years your most likely growing tired of it.

With just looking at the 5 that are on that Arch page it's hard to decided which software to install.

The other thing is what type of head piece with a microphone to use?
W/O even looking (I'll look later)I'm almost willing to bet that there are so many different types of them that it's ridiculous.

Linux Magazine has an article but it doesn't sound like it's complete and I say that because it is said that it's an ambitious project
http://www.linux-magazine.com/Online...ch-Recognition

This sounds like there aren't many Developers working on this-

Quote:

The result is a combination of resources, expertise, and effort that is extremely difficult to organize and sustain in a volunteer project.

Can any of you confirm that any of these: Simon, Speech, Julius and etc. actually work?

Would any of them work with Slackware?

Suggestions are Welcome-:)

frankbell 09-30-2015 09:05 PM

John Kulp, who podcasts at Hacker Public Radio, uses Blather.

If you open the following link and use your browser's search page text feature, you will find several HPR episodes he's done in which he discusses speech recognition:
http://hackerpublicradio.org/corresp...php?hostid=238

Ztcoracat 09-30-2015 09:10 PM

Quote:

Originally Posted by frankbell (Post 5428127)
John Kulp, who podcasts at Hacker Public Radio, uses Blather.

If you open the following link and use your browser's search page text feature, you will find several HPR episodes he's done in which he discusses speech recognition:
http://hackerpublicradio.org/corresp...php?hostid=238

Thanks for the links frankbell-:)

I'll look into the details of Blather:-

jefro 10-01-2015 07:05 PM

I played with viavoice a long, long time ago. The headset was a vast improvement over any other means. Some have noise cancel tech that helps even more.

My moto x has a chip in it to assist with voice and I can say that it works a lot better than apples iphone does. Window phone was maybe equal to the moto x. So, if they can get that to work on a phone, I'd think there is hope of getting it to work on a computer. Opensouce is a different issue.


If I needed this for production then I'd look at Dragon and see if it would run in linux.

Ztcoracat 10-01-2015 07:46 PM

Quote:

Originally Posted by jefro (Post 5428557)
I played with viavoice a long, long time ago. The headset was a vast improvement over any other means. Some have noise cancel tech that helps even more.

My moto x has a chip in it to assist with voice and I can say that it works a lot better than apples iphone does. Window phone was maybe equal to the moto x. So, if they can get that to work on a phone, I'd think there is hope of getting it to work on a computer. Opensouce is a different issue.


If I needed this for production then I'd look at Dragon and see if it would run in linux.

If I'm not mistaken Dragon works only with Microsoft Windows:-

I'll look at headset's this evening and see what I can find.

Thanks for chimming in-;)

Ztcoracat 10-01-2015 07:57 PM

Watching the video that Jon Kulp made was helpful-
https://www.youtube.com/watch?v=gr1FZ2F7KYA

He mentioned that Blather has a lot of dependencies and it was a bit challanging to get them installed but aside from that Blather looks like one of the top 3 in my list.

Blather BTW (didn't know) uses CMU Sphinx as it's backend-

For those of you watching Blather is not in the SlackBuilds repo-

I hope I can find out all of the dependencies w/o going through a painstaking compilation.


-::-Any thoughts on '(finding out the dependencies)' frankbell? <OR> jefro?-::-

jefro 10-02-2015 05:15 PM

I meant Dragon under wine. :)

For dependencies I usually try the build process and see what errors happen. Then fool with trying to get those build and see what errors exist then build those and 4 years later......

Ztcoracat 10-02-2015 07:19 PM

Oh yes, of course wine-
But does wine really work? (I've seen members here complain that it doesn't sometimes)

I'll try building and see what errors I get thanks-:D

jefro 10-02-2015 09:26 PM

Wine has been around a while. Not sure too many things work on it perfectly. Best place to start is at wine hq. https://appdb.winehq.org/objectManag...ation&iId=2077

You can try the other open source projects for sure and see how far you get. Maybe someone else may have build some of the chain already. LQ may have someone that has it all already. Post on new thread maybe errors.

Ztcoracat 10-02-2015 09:34 PM

Quote:

Originally Posted by jefro (Post 5429111)
Wine has been around a while. Not sure too many things work on it perfectly. Best place to start is at wine hq. https://appdb.winehq.org/objectManag...ation&iId=2077

You can try the other open source projects for sure and see how far you get. Maybe someone else may have build some of the chain already. LQ may have someone that has it all already. Post on new thread maybe errors.

Got it- Thanks-;)

Ztcoracat 10-02-2015 09:42 PM

Thank you both for the help and information.

I think I'm going to go with one of the Logitech headsets-
I'll decide later which one I'll purchase later after I either install Blather or CMU Sphinx:-

http://www.logitech.com/en-us/voip-headsets
http://www.newegg.com/Headsets-Acces...at/ID-1080-707

There are other companies like Cyber Acoustics and Plantronics that make headsets but I'm not sure if they work with Linux.

Andy Alt 12-15-2015 09:49 PM

Blather example commands.conf
 
This post is more specific to Blather, but is it okay if I add it to the thread?

I successfully ran Blather on Debian Jessie yesterday. I didn't get the QtGui version to work yet, but did get the Gtk gui operational.

I have contrib and non-free software available in my repos file. All the dependencies of Blather (with the possible exception of what waas needed for the Qt version to work) were satisfied.

I used a git version of Blather. No compiling was necessary, and I didn't "install" it; I used it directly from the source directory.

Here is a prototype of my commands.conf file. It's a bit crude but wanted to do some initial experimentation.

Quote:

#commands are key:value pairs
# key is the sentence to listen for
# value is the command to run when the key is spoken

hello world:echo "hello world"
#programs are opened in the blather terminal window, so the & is needed, otherwise blather stops "listening" until
#the program is closed
open pidgin:pidgin &
open ice dove:icedove &
open ice weasel:iceweasel &
# These lines will print characters of the alphabet to the active window (e.g. a text editor, terminal window).
# Blather had a few problems recognizing certain
# letters, so I used the military phonetic alphabet for a few of them.
# Because blather appends the keyword to the command, a space is needed at the end of each line (xdotool will
# ignore the appended keyword)
# xdotool is available in the Debian repo and is independent from Blather
a:xdotool key a
b:xdotool key b
c:xdotool key c
d:xdotool key d
echo:xdotool key e
f:xdotool key f
guerilla:xdotool key g
h:xdotool key h
i:xdotool key i
j:xdotool key j
k:xdotool key k
l:xdotool key l
mike:xdotool key m
november':xdotool key n
o:xdotool key o
p:xdotool key p
q:xdotool key q
r:xdotool key r
sierra:xdotool key s
t:xdotool key t
u:xdotool key u
victor:xdotool key v
w:xdotool key w
x:xdotool key x
y:xdotool key y
zulu:xdotool key z
click left:xdotool click 1
click right:xdotool click 3
click middle:xdotool click 2


All times are GMT -5. The time now is 09:00 AM.