Okay so, what is the Holy Grail for speech recognition? I would submit to you that when the artificial intelligent computers, perhaps coupled to supercomputers can spot a particular voice in a stadium, or someplace like Grand Central Station in the middle of the day, that they will have reached the pinnacle. Do you think this is impossible? Because I don’t, in fact as I look at FRT or facial recognition technologies, I realize that we are already at that point.
That is to say that FRT has advanced so far that they can now spot an individual at a busy airport, or all the individuals at a busy airport, and then go on to a website or social network like Facebook, and determine who everyone is, and match their faces with their identities, even if their faces are not listed in the database used by the security software. That is amazing. Equally interesting, is the fact that every human voice has idiosyncrasies much like a fingerprint that are unique to that individual, that uniqueness can give away the person talking.
However, currently speech recognition technologies are not to the point that they can filter out many voices in a large group, party, or an abundance of anonymity surveillance background noise to the point where they can isolate each individual noise, and separated out the whole. It gets even more complicated the more voices, sounds, and noise which permeates the background. Nevertheless, I don’t believe it’s impossible, and I believe that someday we will get there.
This type of Holy Grail for voice-recognition software is called; Global Speech Recognition. The ability to go anywhere, regardless of the interference of other voices and noise, and identify the individual by their voice patterns within the crowd. It takes a lot of work to find the patterns of each individual voice, and then separate that out, but I would submit to you it is not impossible, and therefore because it can be done, it will be done. Perhaps in a busy coffee shop it might be possible.
This technology has obvious advantages for surveillance, security, and all sorts of other things. There may be challenges and questions about privacy in the future as this technology advances as well. Who knows, perhaps such advances already exists with regards to voice-recognition, and soon it will be possible to isolate hundreds if not thousands of other sounds, patterns, and voices to find the specific individual through this recognition technology in the future. Indeed I hope you will please consider all this and think.