Hello!

It's me again :)

Today I released the first beta of simon ever. Version 0.2-beta-1.

The 0.2 series of simon introduces a LOT of new features:

• Porting to KDE4
  
• Model synchronization with automatic backup
• Support for special characters and numbers
  
• Support for SPHINX dictonaries
• Support for PLS dictonaries
• Network audio-streaming to the server
• "Live" command-plugin loading and un-loading
• Julius integration; Including automatic re-initalization of the recognition when the model changes
  
• Better sound configuration (also effects the recognition now)
• And a lot more...

simon 0.2 beta 1 is available as source tarball, ubuntu debian package and windows installer from http://sourceforge.net/projects/speech2text

I would love to get some feedback!

Also, I set up a blog on blogspot (http://simon-listens.blogspot.com/) so that everyone that is interested in simon can follow the development more easily (RSS feed: http://simon-listens.blogspot.com/feeds/posts/default)

regards,
Peter

--- (Edited on 12/27/2008 6:34 pm [GMT-0600] by bedahr) ---

Hello Voxforge Community!

The first stable simon version is born: simon 0.2 has been released.

The full announcement can be read on our blog.

--- (Edited on 7/10/2009 2:06 pm [GMT-0500] by bedahr) ---

HI bedahr,

Well done!

Ken

--- (Edited on 7/13/2009 11:48 am [GMT-0400] by kmaclean) ---

Yeah, well done, bedahr. I was able to recognize 20 words correctly after importing the German PLS dictionary.

Ken, do you have any idea how the Voxforge acoustic models (English, German) could be imported into (a future version of) simon? Or it would be interesting to import some of the Voxforge prompts (wav/transcriptions) into simon. But how could I achieve that?

--- (Edited on 2009-07-16 1:23 am [GMT-0500] by ralfherzog) ---

Hi Ralf,

>Ken, do you have any idea how the Voxforge acoustic models (English,

>German) could be imported into (a future version of) simon?

I am assuming that since it uses HTK format acoustic models, you should be able to just replace the hmmdefs, macros and tiedlist files with  VoxForge's versions of these files.

>Or it would be interesting to import some of the Voxforge prompts

>(wav/transcriptions) into simon. But how could I achieve that?

I don't know the internals of Simon.  It might be as easy as adding a few entries to a text file, or it might use XML or some database format.  Peter from the Simon project would know.

Ken

--- (Edited on 7/20/2009 12:42 pm [GMT-0400] by kmaclean) ---

Hi Ralf! Hi Ken!

Ken is right: replacing the model files in ~/.kde/share/apps/simond/models/<your user>/active with the voxforge model files will work (I actually already tested that a while ago).

However, with simon 0.3 we are going to make it even easier by providing a gui for that (possibly with the option of automatically downloading a recent voxforge snapshot) and also adding a feature to adapt a general model (like the voxforge one) to your voice instead of generating an entirely new one. Of course all that will be configurable and no functionality will be lost after the introduction of this new feature.

On the prompts issue: simons philosophy has always been to be as thin a layer as practical. That means where possible we will not introduce new file formats if there are no serious reprocussions. In case of the prompts file we really use just that: plain text prompts file as the HTK expects them. You can of course add samples to them but make sure you place them in the configured samples folder so they will be found during the compiling of the model. The prompts file is located at ~/.kde/share/apps/simon/model/prompts. However, in order for the synchronization to be triggered you would also have to manually update the "TrainingsDate" value to the current date/time in the file: ~/.kde/share/apps/simon/model/modelsrcrc but that sounds harder than it is.

Greetings,

Peter

--- (Edited on 7/20/2009 12:44 pm [GMT-0500] by bedahr) ---