(PhysOrg.com) -- Microsoft recently obtained a patent designed to create an Automatic Censorship of Audio Data For Broadcast . The invention is intended to act as a filter for live broadcasts where it is impracticable to delete or make inaudible certain undesired words or phrases. Additionally, other audio streams like music or games can utilize the automatic censor.
An input audio data stream comprising speech is processed by an automatic censoring filter in either real time mode or batch mode, producing censored speech that has been altered so the undesired words or phrases are either unintelligible or inaudible. The substituted speech is not bleeped out, but modified in the speakers own voice.
According to the patent abstract, the reason for the invention is that it is common for radio and and television broadcasts of live events to be delayed a few seconds before the audio data is transmitted to enable real-time (censors) content reviewers to evaluate whether the content contains objectionable language. Broadcast television and radio are heeding the demand by the public to essentially "clean up" the airways for the general public and children in particular. The degree of censorship depends on the targeted audience and the nature of the event. Specifically, certain words which are commonly referred to as profanity, obscenities and sexually explicit words are targeted by censors to be unacceptable for public broadcasting.
In certain situations where there are multiple individuals speaking as in a game-play scenario it is almost impossible for a human censor to catch all extraneous remarks or words. This is where the Automatic Censorship of Audio Data for Broadcast fills in the gaps. Instead of bleeping out objectionable language, the patent will essentially supply a substitute phrase or word or make the word inaudible to the listener. While the patent can employ the "bleeping" of objectionable words, it will attempt to overwrite the objectionable content with a more acceptable term or word. The acceptable word or phrase is created by using previously uttered phonemes of the speaker that can be combined to produce the more acceptable version. Thus, the censored content appears and sounds like the speaker´s own words.
The Automatic Censorship of Audio Data for Broadcast patent is not without its critics. On-line forums and bloggers are not uniformly impressed with the patent. According to one comment on Slashdot, "Come on, if you´re old enough to play the game, you´re old enough to either deal with it or tell them to stuff a sock in it." Other comments include the impossible task of coding various types of a new breed of curse words. To wit, another commentator congratulates Microsoft for expanding the lexicon of profanity to include a more creative style of profanity and obscenity in order to fool the program.
One thing is for sure, the Automatic Censorship of Audio Data for Broadcast evokes a discussion on the subject of censorship. Comments are welcome on this post without %$ included, if at all possible.
Explore further: Peer-review science is taking off on Twitter, but who is tweeting what and why?