Normalizing audio?

[yetanotherlogin]

Hello

I have an audio stream of a conference where sound level varies greatly because speakers didn't speak with the same strength.

Contrary to what I expected, Audacity's Effect > Normalize didn't edit the stream so that audio volume would be the same throughout the file.

Does someone know of a way to homogenize the sound level in an audio stream?

Thank you.

[Mr Chris]

Normalizing brings up the volume of everything by an equal amount, so everything remains in proportion to the original. The effect you need is compression - that much I can help you with - but I'll have to let someone more qualified teach you how to apply it well.

[Djard]

Use the Compressor in Audacity or Thomas Mundt's "LoudMax" plugin. The two main settings in the Compressor are "Threshold" and "Ratio." Set the former to -18 dB or more and the ratio to about 6:1. The higher the "Threshold, the greater the reduction of dynamic range (DR). The "Ratio" setting does not reduce DR as much as the Threshold setting. There's an excellent online step-guide: just search for "Audacity101." You may need to click twice on the search icon in your browser to bring it up first. Boosting the 4K frequency by 2-3 dB with the EQ might help in your project.

[yetanotherlogin]

Thanks much for the infos.

[hello_hello]

Or try the Levelator.
http://web.archive.org/web/20130729204551id_/http://www.conversationsnetwork.org/levelator/

[Djard]

If the recording is of great value to you, the most effective way by far to correct the problem you describe is to manually amplify the weak sections, which task is also the most time-consuming. You would use the "Envelope" tool in Audacity, highlighting the weak areas of the waveform, clicking before, mid and after each weak wave then dragging the mid point up to increase the amplitude. In this manner, you will have the maximum amount of control over the project and render a near-perfect recording. Again, the steps are given in the online "Audacity101" guide.