Microsoft Patents Real-Time Profanity Filter
It's no secret that most radio talk shows have a slight delay before a caller's audio reaches the air. The obvious reason: to make sure no "f-bombs" reach the sensitive ears of the public --- or the FCC.
Well, Microsoft has been awarded a patent that might eliminate the need for that delay. The patent, 7437290, is for "Automatic censorship of audio data for broadcast" was first filed in 2004. Here's how it's described:
In other words, if the filter matches a word to an undesirable word, it'll bleep it. It has applications in more than just radio, though. You could imagine it being used in video games, in perhaps a "family mode" in which any profanity uttered via voice chat is bleeped.
We could also imagine, not now, but with an enhanced form of this application, it being used to filter voice messages of any sort that might fall under censorship rules in such totalitarian countries as Burma or China. But that would require the filter to not just "listen" for words, but for phrases or even whole sentences, so that will take some time to develop.
OK, next step: some sort of video filtering to prevent a repeat of the Janet Jackson Super Bowl "accident."
Well, Microsoft has been awarded a patent that might eliminate the need for that delay. The patent, 7437290, is for "Automatic censorship of audio data for broadcast" was first filed in 2004. Here's how it's described:
An input audio data stream comprising speech is processed by an automatic censoring filter in either a real-time mode, or a batch mode, producing censored speech that has been altered so that undesired words or phrases are either unintelligible or inaudible. The automatic censoring filter employs a lattice comprising either phonemes and/or words derived from phonemes for comparison against corresponding phonemes or words included in undesired speech data. If the probability that a phoneme or word in the input audio data stream matches a corresponding phoneme or word in the undesired speech data is greater than a probability threshold, the input audio data stream is altered so that the undesired word or a phrase comprising a plurality of such words is unintelligible or inaudible. The censored speech can either be stored or made available to an audience in real-time.
In other words, if the filter matches a word to an undesirable word, it'll bleep it. It has applications in more than just radio, though. You could imagine it being used in video games, in perhaps a "family mode" in which any profanity uttered via voice chat is bleeped. We could also imagine, not now, but with an enhanced form of this application, it being used to filter voice messages of any sort that might fall under censorship rules in such totalitarian countries as Burma or China. But that would require the filter to not just "listen" for words, but for phrases or even whole sentences, so that will take some time to develop.
OK, next step: some sort of video filtering to prevent a repeat of the Janet Jackson Super Bowl "accident."
