While technology has been "speaking" for decades now, it's only recently that speech has stopped being a novelty act and has become a legitimate interface. But for that transition to happen, computerized speech itself had to improve dramatically from the days of heavily digitized, halting sounds that approximate language (think of all those computers from the 1980s spitting out, "I. Am. A. Com.pu.ter.").
The video above shows how much work goes into creating a computerized voice that's natural enough to interact with device users. It is, as one might imagine, both interesting and complicated, as it's not just capturing accurate sounds, but accurately portraying the fluidity between the sounds. For example, an "ah" sound has slightly different inflections based on what sounds come immediately before and after it, so multiple intonations need to be recorded. Then there are the added challenges of speech-to-text technology, both in terms of turning a user's speech into typed text and having the device "read" typed text out loud to the user.
Partway through the video, you'll hear a very familiar voice. It belongs to Allison Dufty, but to the tech world, it's all Siri, the iPhone's voice-assistant feature. If you've ever wanted to put a face to that voice on your phone, now you can.
This story, "Video: Siri gains its voice," was originally published at InfoWorld.com. Keep up with the latest tech videos with the InfoTube blog. For the latest developments in business technology news, follow InfoWorld.com on Twitter.