Taperssection.com

Gear / Technical Help => Post-Processing, Computer / Streaming / Internet Devices & Related Activity => Topic started by: vanark on May 26, 2009, 11:30:44 PM

Title: Audacity - Hiss Removal
Post by: vanark on May 26, 2009, 11:30:44 PM
I received a recording of an acoustic string trio with a little harmonica, kazoo, jawbone of an ass and plain old stompin thrown in.  The recording was recorded with peaks only at about -15db.  Even without bringing the levels up, the hiss was very noticeable.  Bringing the levels up 12 db to -3.0 peak brought all the hiss with it.  Using the Noise Removal tool, I found a pretty clean noise sample and went at it.  Listenable now, I think.  Comments?

Audacity v1.3.6 beta

sample_hiss.flac
http://www.sendspace.com/file/k104uv (http://www.sendspace.com/file/k104uv)

sample_dehiss.flac
http://www.sendspace.com/file/i41dpu (http://www.sendspace.com/file/i41dpu)

edit: new SS links.
Title: Re: Audacity - Hiss Removal
Post by: Brian Skalinder on May 26, 2009, 11:36:03 PM
Not sure if it's just me, but attempting to download via SS gives an error:  The download link has been updated (error: #10).
You may now download the file.
  I try again, and the same error repeats.

Title: Re: Audacity - Hiss Removal
Post by: vanark on May 26, 2009, 11:50:59 PM
Thanks for the heads up, Brian.  I deleted the old links and created new ones.  I tested them and they work for me now.
Title: Re: Audacity - Hiss Removal
Post by: bbernardini on May 27, 2009, 09:44:14 AM
Another option for hiss removal is to find the approximate frequency of the hiss using the "Plot Spectrum" option. There's a tool under "Effect" (possibly vocal removal) that will allow you to specify a frequency range for removal. I've found this to be extremely useful at removing both hiss and buzz without any drastic effect to the rest of the recording.
Title: Re: Audacity - Hiss Removal
Post by: vanark on May 27, 2009, 09:54:38 AM
I think the Noise removal did an adequate job in this case.  Anyone have any thoughts the sound of the samples?

I think because I was able to find a 2 second sample of almost 100% hiss, the utility worked better than most have experienced.