Voice recognition is one of the most attractive thing that a smartphone package offers. It can’t be more interesting, the user has a query for the virtual assistant Siri, Alexa, or Google whichever it may be it related to anything under the sky. In a handful of milliseconds, the query returned mostly correctly interpreted. Voila!
However, the entire phenomenon has flaws too. The time lag factor in such a phenomenon can be a test of patience.
This is because voice from the user’s smartphone travels to the virtual assistant’s server that can take a handful of milliseconds to multiple full seconds. This can be even longer if the packets get lost in the ether.
Despite Limitations, Google’s Voice Recognition by phones a step forward
What can be best to prevent such scenarios? To do voice recognition on the device itself. Adding this capability would be the best thing for companies that offer virtual assistant services. This would, however, involve considerable computing power.
It is not that simple as it sounds. To hear a sound word by word, and understand it correct involves a lot of context about language and intention. Google’s work in this space has come to a fruition.
But Google’s offering has limitations too. One, it works only in Gboard- Google’s keyboard app. Two, it only works on Pixels and in American English. The pilot is kind of a stress test before the real thing.
Elaborating on the development, Google wrote how trends in the industry can help make it mainstream. As specialized hardware and algorithmic improvements converge, techniques that currently have small use can soon be adopted for more languages. The domains of application may broaden as well.