Made structural update that now makes the dictation be more robust in terms of: model, mode, and result. This way there are a more deterministic intuition behind what dictation run is actually run (what settings was included). The result is a more responsive application that also have better way of handeling the states of the different buttons and modes available for specific languages and models.