Qt5.x IVI Speech Recognition

[Peter Grasch]

On 24 Apr 2014, at 08:13, Ramakanthreddy Kesireddy <Ramakanthreddy.Kesireddy at techmahindra.com> wrote:

> Hi Peter,
> 
> Am finding it difficult to create speech model.
> As you said, you can build a speech model with Simon and export it.
> 
> Could you please let me know how to create speech model for Julius+HTK as we would use static dictionary?
Roughly, these are the steps you will want to follow:
* Likely you will not want to build your own acoustic model so look for / select the correct base model (if you want to use HTK / Julius, use the released [EN/VF/JHTK] Voxforge Binary).
* Import the correct shadow dictionary for the selected base model (the english voxforge dictionary available from voxforge.org for the above named base model)
* Create a new scenario and assign it the relevant task vocabulary and grammar.
* At this point, Simon will automatically create your speech model in the background. Once it’s done, you can export it from the model configuration page in the application preferences.

You will find more detailed instructions about the individual steps in the Simon user manual.

Best regards,
Peter