Delete sound higher than specific DB.

I have a short sound recording of a music event that i made with my smartphone. I’ve converted it to a wav and mp3 file. Cant seem to figure out how to delete some loud intermitant talking by a person next to me. How do i delete sound over x DB or remove that specific voice. I know I can clip out those parts but thats not optimal. i have Version 2.0.5.

And the only other option won’t be optimal, either, where you cut a silent hole in the show.

You can’t split a musical performance into individual instruments and performers, and that talker is now a performer in your recording.


Isn’t there a way to set cut off sounds above a specific decibel?

I’m trying to understand exactly what you hope to do.
What do you mean by “Delete”?

Why is that not optimal?

The music is in a steady lower decibel range and the voice spikes are much higher. I was hoping to limit the track to a lower decibel range thereby cutting off the voices that are at a high decibel. Cutting out the voices will obviously cut out the music as well. I would rather not have any sound above a specific DB with some hint of those voices probably still remaining in the lower range. Thanks

I’m still not quite sure what you mean.

1)Do you mean that you want to reduce the level when there is loud talking so that it becomes silent?


  1. Do you want to “delete” the loud sounds, so for example, if you have 10 seconds of quiet sound, followed by 3 seconds of loud talking, followed by 10 seconds of quiet sound. After deleting the loud talking you will have 20 seconds of quiet sound.

Something along the lines of the 1st choice you listed. I’m trying my best to explain, if I look at the sound track and the music is very steady and even at a slightly fluctuating decibel level and then you see a spike on the track where someone talked loudly…I want to eliminate that spike, essentially chopping off that mountain peak and only have sounds at a max decibel level just a little higher than the music. Is it possible to do this?
I thought of trying it with the voice removal feature but I have no idea how to identify what frequency the voices are at to remove them.

Try “Chris’s Dynamic Compressor”. It will even out the volume changes.
Try it with all default settings except for the “Ratio” (the first control). Try setting the Ratio to about 0.8
The plug-in is available here:
Instructions for installing:

Hello everyone.

I have same question as thread starter but I will try to explain my question clearer.

I have audio record, Audacity can break it into frequencies with certain amplitudes using Fast Fourier Transformation (FFT), how do I ask Audacity to remove frequencies from FFT result hitting (for example) 0.70 loudness and above?

I don’t know of any plug-in that will do that but you can try [u]Spectral Editing[/u].

“LEVELS” plugin can do that, see …