This scenario requires you to have a working installation of Perl available as '/usr/bin/perl', with the Tk module installed ('sudo cpan install Tk' should do this for you). You also need to ensure you have the Perl Net:
Bus module installed and this requires a few other modules. Normally, using cpan to install DBus should automatically install all prerequisites, but my system (Fedora 14) required me to install the Perl XML::SAX module separately, before attempting Net:Bus. On my system, the following three commands got everything ready for the script:[indent][doug@machine]$ sudo cpan install Tk[doug@machine]$ sudo yum install perl-XML-SAX.noarch
[doug@machine]$ sudo cpan install Net:
Bus[/indent]You will need to adapt these commands to suit your system environment, user permissions and version of Linux.The scenario makes calls to a Perl script called dbus_marble.pl, which must be executable by you and in your path. You need to download the script from https://docs.google.com/leaf?id=0B7ed9W7zAS2XODBjNjNiYTMtNWQxZS00NTg0LTkzZDEtOWE5MjlkNTkyMDQ2&hl=en&authkey=CJXS7P4K and install it appropriately. On my system, I have made the script executable by everyone ('chmod 755 dbus_marble.pl') and created a symbolic link to it in /usr/local/bin. You should do whatever is prudent and possible on your system, to achieve the same end.
It is intended that users install at least two other scenarios to get full control of the application without using a keyboard, other than for text entry. The two scenarios you should install are:[indent]* [EN/VF] Menu Navigation - for menu accelerators
* [EN/VF] Marble - for non-DBus Marble commands[/indent]
This scenario has two training texts:Commands, containing all commandsNo Prefix, which omits the Marble prefix, to speed up training
This scenario has been tested with the VoxForge static model.
YOU can help to improve the static model by donating voice samples to VoxForge, at http://www.voxforge.org/home. VoxForge is building publicly-available speech models and depends upon public donations of voice samples to improve the model. Please go to their web page and follow the links to 'Submit Speech Using Computer'. All you need are a microphone and Java enabled on your browser: there are helpful links if you have any trouble.
Post any questions or problems on the Help forum at https://sourceforge.net/projects/speech2text/forums/forum/672427.
Ratings & Comments
0 Comments