SpokenSystemAndVoiceCommand

Differences between revisions 11 and 12
Revision 11 as of 2007-11-21 14:40:32
Size: 3047
Editor: i59F71AB2
Comment:
Revision 12 as of 2007-11-21 15:08:04
Size: 3323
Editor: pat-address
Comment: comments
Deletions are marked like this. Additions are marked like this.
Line 56: Line 56:
 * Regarding voice recognition, gnome-voice-control was a SoC 2007 project, is packaged and functionnal in Gutsy. An equivalent solution is integrated into KDE.
 * What would be awesome is to have early activation of the system so that we can use it during on login screen

Specification

Summary

As Ubuntu wants to be a serious alternative to M$ systems, it would be useful and interesting to provide a spoken system (which is complementary of the spoken boot) and to be able to launch current tasks using our voice.

Rationale

These last years, a very important effort is done to ameliorate the user interface. In order to perpetuate and revolutionize it, we should be able to innovate, and try to include a spoken system and at least a few voice commands in order to make the users life different, and provide something that is unique and ambitious.

Use cases

  • Nicolas wants to be able to communicate more naturally with his system, hands free, and he wants his system to read for him the main outputs it feeds back.
  • Paul, a blind person, can also benefit from this system, with a more interactive OS.
  • It will be great if we have a multilingual voice controlled system

Scope

The spec aims to provide an advanced stack of tools preconfigured for doing voice control of the system and for the system to read the error and information messages. The voice control could be made of a voice recording assistant which, a bit like the "keyboard shortcuts" one, can associate a word with a system command (instead of keys combinations) but instead of it, tolerates a few differences, because voice is not always the same. The information and error messages reading could be done using espeak, festival or similar, like the system already included in kde (kttsd, ksayit and kmouth with a voice synthesis system)

Design

  • Use packages that are already in our repositories, like espeak and festival, and try to find those which can be useful for the voice command system part.
  • Maybe use improvements and/or additional packages or programs provided by developers interested by this challenge.

Implementation

* Not started


Current Status

  • Initial Drafting

Things being worked on / problems

Things feasible for Gutsy release

Things yet to come

Improvements

Comments

  • Document structure inspired by the [:Spec/FeistyTelepathy] spec wiki page, thanks to you. --kalon33
  • ["Warbo"]: The only advanced Open Source voice recognition software I know of is Sphinx, which Google tells me is based here: http://cmusphinx.sourceforge.net/html/cmusphinx.php Hope it helps Smile :)

  • This also ties to the disability accomidation aspect of the product. [http://www-304.ibm.com/jct03001c/software/pervasive/embedded_viavoice/ IBM Viavoice] has been bandied about as available technology for FOSS use. Maybe some nice person from IBM could come to the Developers Conference in October and explain this to us.

  • Regarding voice recognition, gnome-voice-control was a SoC 2007 project, is packaged and functionnal in Gutsy. An equivalent solution is integrated into KDE.
  • What would be awesome is to have early activation of the system so that we can use it during on login screen


[:CategorySpec]

SpokenSystemAndVoiceCommand (last edited 2008-08-06 16:13:55 by localhost)