|
| Link: http://simon-listens.org
Depends on: KDE 4.x
Downloads: 2336
| Submitted: Jan 18 2008
Updated: Jul 10 2009
| |
Description:
The project provides a ready-to-use interface for the julius speech recognition engine. It integrates into X11 and Windows.
simon can import dictionaries, grammar structures and compile a language model with the help of the HTK toolkit of the university of cambridge.
With the recognition results it can type texts, simulate shortcut sequences, start programs control your mouse and more.
simon comes with a complete manual describing the system. This manual and setup instructions are also available from our wiki: http://simon-listens.org/wiki
Please don't hesitate to contact me if you need help!
Changelog:
0.2:
Fixed the remaining known issues and added the manual.
0.2 rc1:
We've done a _lot_ of testing. Basically we tested every feature as well as we could and came up with a lot of bugs.
Well those are all fixed now :)
There are two issues left which are sadly quite bad: One crashes simon and one causes the recognition to stop without any error or warning what so ever. Both only get triggered when re-initializing the recognition and are related to Juliuslib. I hope to get that fixed soon.
But all in all, simon should be quite stable and ready for some productive use!
0.2-beta-3:
Introducing three new command plug-ins:
* List
A list of other simon commands; The user selects one of the commands using numbers 1-9; If there are more than 8 commands in the list, the list only goes from 1-8 and 0 and 9 are can be used for scrolling (0: up; 9: down).
* Composite
Combines other simon commands in one "Macro". The user can even define delays between the invocation of two sub-commands.
* Input Numbers
A calculator-like front-end to input long numbers with ease.
Also, the desktop-grid can now be completely controlled by voice.
The recording widget now displays the current "loudness" of the input while recording. This should help to weed out not working / badly configured microphones.
Updated Julius to the current SVN versions with minimal modifications.
Bug-Fixing all around - especially related to sound and recognition.
0.2-beta-2:
Fixed various bugs related to the grammar which caused it to not get saved most of the time. It is now again possible to add multiple words at once without losing samples. All input files to the importing functions (dictionary, texts, etc.) are now checked if they exist before the importing begins. simon now compiles also on Ubuntu Hardy Heron (KDE 4.0) Removed KDE-Includes when building julius (a lot of warnings due to strict compiler flags) All libraries are now installed correctly.
0.2-beta-1:
Bugfixing all around.
Fixed a whole lot of libjulius / sent related issues regarding stopping / starting of the sound-stream.
Fixing the Windows Version.
Introducing CPack support for quick package building.
0.2-alpha-1:
Porting to KDE 4
Removed Juliusd in favor of simond:
* Network audio streaming
* Recognition and model compilation on server
* No processor intensive calculations on the client (opens the door to development of solutions for embedded devices, etc.)
Revised and greatly improved command plugin architecture
* Individual triggers per plugin
* Runtime module loading / unloading (even the configuration of the modules)
Introduction of ksimond - a frontend to the simond recognition server
And much, much more...
0.1-alpha-2.1:
A lot of compatibility fixes.
0.1-alpha-2:
Too many changes to count - here the most important ones:
Fixed about a gazillion bugs.
Streamlined the interface.
The command-system has been completely rewritten.
A few performance optimizations.
Introduced a structured build-system.
Switched to portaudio backend for more compatibility.
Did I mention fixing bugs? Lots of them...
0.1-alpha-1:
Initial release
License: GPL
|
| |
add commentback
|
bad
good
slashdot it
digg it
del.icio.us
