Speech Recognition, Virtual Assistance, Future Phone

Discussing the speech recognition technology will be a long and technical writing. While the engineers discussing it, let's talk a little lighter, but no less interesting.

As part of artificial intelligence, voice recognition systems require special programming and can not be done by coding each task one by one. The core of the personal assistant, like Siri, is in voice processing. If the voice has been processed correctly and accurately, the results can be forwarded to data processing machines. Related to this, Google actually has a richer source.


The process that occurs in a virtual personal assistant phones more or less like this:
1. The voice coming through the microphone of  users' phone
2. The voice is processed by the speech recognition engine
3. Results forwarded to data processing machines, involving databases.
4. The results are then sent to mobile users in a certain format
5. Phone translates the results according to preferences of mobile users, whether voice or text.

I have ever tried, Vlingo, appears to perform step 2 on the server side, so that the phone needs to transmit data in voice format. For phones with slow Internet access, this does not work. Problem of network errors most often occur. Ideally, the second process occurs in the mobile phone users, so that data transfer can be made in a minimum amount over the network.

The step 2 is the advantages of Apples's Siri. Not ideal if you compare the Vlingo and Siri. Apple bought Siri with the price of 200 million dollars, not cheap for companies that do not have the vision to use it. But for of Apple, it's a good deal. It is still discussed everywhere and become selling point of Apple iPhone 4s.

Some people think that talking to phone feel weird and awkward, but naturally, it happen at the first time we use new things when another don't. When i use my first cellphone, it feel weird too when many people staring at  me while talking using a big brick old cellphone.