Monday, February 8, 2010

User Guided Audio Selection from Complex Sound Mixtures

by Paris Smaragdis

This article disscussed the use of complex sound, mainly human voice, to filter out specific sound in audio recordings for editing or removal. The concept behind the idea is that a user mimics that sound which they would like to work with and then the application filters out that sound. This can be used for such things as removing a singers voice from a track for those who wish to be able to hear just the background music. This is not a new concept to sound editing however the ability to use a specified sound "object" input is new to this field.

I believe that this is an awesome idea that could very well revolutionize normal user sound editing and mixing. However some faults I see with this work is the ability to truely mimic only the soudns that you would like to remove may limit your ability to properly filter the audio file.


1 comment:

  1. I see your point on the ability to mimic. I've known some people who are completely tone deaf. I'm not sure what they're using to detect the intended part, but I wonder if a tap on a surface that corresponds to sounds changing would be an alternative.

    Did they say anything about getting things like drum parts?

    ReplyDelete