Hello, I’m trying to extract some sound from a mixed sample and was wondering if there is any technique or plugin that could analyze one sample and remove it’s exact frequencies from another, or something to this effect? What I have is about 30 seconds of speech that extends a few seconds into the beginning of a piece of music and I’d like to remove it. The mono technique won’t work because that part of the music is also mono and I’ve tried a vocal removal plugin which worked by removing an entire frequency range which didn’t work well
See this topic: http://audacityteam.org/forum/viewtopic.php?f=26&t=4809&p=18927#p18927
Don’t expect it to work well, but you may be lucky.
So the DJ talked right over the music you wanted? Darn those people.
The exact process you asked for is how the Noise Reducer works. You play samples of the announcer and select them for the noise reducer to derive a profile. Then you apply the tool to the rest of the show and It tries to subtract all of those frequencies from the performance.
Don’t even try this in anything earlier than 1.3.3. The noise reducer in 1.2 was terrible.
I don’t expect it to work because human voices have too many common frequencies with music. But it will be fun to mess with. Now if you wanted to subtract an air conditioner from music, we got that one licked.
koz