Progress updates for Voice Control 2.2.0
It’s been a month and half since last update, and I would like to share some updates on progress towards the version 2.2.0.
These weeks I’m constantly working on the integration with the offline recognition engine, and finally these days I was able to make it works with words that don’t exist in the engine pronunciation dictionary (it is mainly needed for contacts, songs and applications names).
The drawback of this is that I had to integrate a “Pronunciation Generator Engine”, that takes about 6 seconds to startup, and adding them to the 4 seconds required by the offline recognition engine, it raises the application total startup time in offline mode to 10 seconds on a Nexus S.
But there are some good news:
- Firstly, it is a lot less than the time required by the original “Pronunciation Generator Engine” integrated (it was more than 3 minutes!)
- Secondly, and most importantly, it is not required on every application startup, so it will be used only the first time that the offline recognition is used (or after some change to the phone data that requries the generation of new pronunciations).
So, right now, it is an acceptable drawback compared to not be able to use the offline mode at all.
There are still some other things to complete and fix, but I hope to be able to release a working beta for English, French and Russian beta testers in the next few days, in order to finally check how it performs “on the road”.
However, at this stage, the “offline mode” is quite far from perfect, so I’m thinking to provide it initially as an “experimental feature”, in order to allows users to evaluete and use it only if it works. This will allows me to work on other pending features for the next version.