Google Voice search is now better and more accurate. Now, it will offer better results even in noisy environment. Today, Google announced refinement on its neural network-powered voice recognition system that was released in 2011. It has been slowly improving ever since and now Google has added better neural network acoustic models using Confectionist Temporal Classification (CTC) and sequence discriminative training techniques.
As we know, Google is constantly working on its services and refining them to a new level since beginning. With this improvement, one can use voice search more accurately even in the noisy surroundings and also in real time it can perform necessary analysis. This new acoustic model is more accurate and faster to respond to voice search queries.
As per Google blog post, “Previously speech recognizer, split the words into small frames of 10 milliseconds of audio and then each frame get analysed. If the user speaks the word “museum” for example – /m j u z i @ m/ in phonetic notation – it may be hard to tell where the /j/ sound ends and where the /u/ starts, but in truth the recognizer doesn’t care where exactly that transition happens: All it cares about is that these sounds were spoken.”
“Our improved acoustic models relay on Recurrent Neural Networks (RNN). RNNs have feedback loops in their topology, allowing them to model temporal dependencies: when the user speaks /u/ in the previous example, their articulatory apparatus is coming from a /j/ sound and from an /m/ sound before. Try saying it out loud – “museum” – it flows very naturally in one breath, and RNNs can capture that,” It added.
This new acoustic models are already being used on android and iOS for voice searches and commands in Google App. So try this accurate, noise free and faster voice searching. On your Android device, say “OK Google” to search.
Source: Google Research Blog
