I think it's also inaccurate to say the Nexidia technology turns audio into text. It doesn't. That's basically what Adobe's speech-to-text function does and it generally doesn't work very well.
Nexidia analyzes waveform shapes, with audio split into some sort of phonetic or syllabic divisions. These waveform patterns are matched against a known library for a given language. The result is a waveform match of recorded audio against the waveform for the existing text entered by the user. It's not turning that audio into text. That's why spelling isn't critical as long as an incorrect spelling still sounds the same as the correct spelling would be pronounced.
Also Soundbite is Nexidia's own product with a few hooks added by Boris, along with marketing. So less overhead, I presume. And they have control over how they want to handle the licensing. It's basically derivative of the bigger search tools they sell to enterprises.
- Oliver
__._,_.___
Posted by: oliverpeters@oliverpeters.com
| Reply via web post | • | Reply to sender | • | Reply to group | • | Start a New Topic | • | Messages in this topic (16) |
.
__,_._,___
No comments:
Post a Comment