AI Could Soon Give Speakers Directional Voice Detection

By this point the concept of a device being able to recognize your voice and understand whatever it is that you might be saying is the sort of thing that is rather common if you think about it. However, the direction from which you are speaking is the sort of thing that technology has not really managed to pick up yet for a lot of different reasons. This means that a machine would need what is called a “wake word” to be able to tell whether you are giving it some kind of a command or if you are just speaking to someone else.

A new AI that is being developed by Carnegie Mellon University might just make it so that wake words would no longer be required for people that are looking to voice activate anything at all. This is because of the fact that this AI would have the ability to detect the direction in which your voice is coming from, thereby giving it a better understanding of whether or not you are trying to give it a command without requiring a specific word to make it realize that you are speaking to it.

This is a pretty huge step in terms of ensuring that everyone is able to use voice detection technology naturally rather than using awkward phrases to get things to work. It’s all about making things smooth and seamless after all, and this research can help technology go a long way towards achieving this kind of goal. While it might not revolutionize voice detection just yet, after all this is just initial research and there will be quite some time before it can be rolled out to the average consumer, but it is still a major step that people will be talking about due to its long term relevance.



Read next: Facebook’s AR Glasses may supplement the hearing aids of hearing-impaired people in the future

No comments:

Post a Comment