The remote part of the audience was sitting around a table with a single mike in the middle, and the boardroom's acoustic was terrible. I clearly lost a part of the conversation, and several times had to ask a speaker to repeat.
Could the following be a solution that we'll see in a not so distant future?
Spoken words can be converted to text locally, with a text, not a digitized audio, being sent across a network. On a receiving side such text can be converted back to speech.
Speaker's voice pitch will probably be sacrificed to large extent. Intonations, and some subtle parts of speech will be lost, including 'um', 'er', or 'ah'. Some additional data, e.g. voice pitch and tempo, need to be transferred along with a plain text. Everyone will sound a bit as Stephen Hawking.
While this approach will hardly go well with medium to high fidelity conversations, it will still suit a large pool of situations. Also this will allow a language translation facility to be placed anywhere along the conversation path: at sender, at receiver, or at carrier.
Skype Translator
http://www.cbsnews.com/news/how-skype-is-becoming-like-star-trek/
No comments:
Post a Comment