This is probably an extremely easy question for you, but I'm very new to this. I found a video file of my niece talking and screaming on my phone, but I'm trying to hear the voice in the background. Both voices are talking at almost the same time. The video file is about 40sec long. I was playing around with AVS Audio Editor but couldn't figure out how to lower the baby's voice, if it's even possible. It's the high frequency that I need to lower, the other is a low frequency, thinking that it can be done? Like I said, I'm new to this, so I got very confused with the terminology on other sites and forums. I'm hoping you guys can help me out or point me in the right direction at least.
Normally mixing is done with a single voice on each channel because it is impossible to completely separate one audio signal from another after they are combined. Sound is additive and there isn't a way to identify one from the other to pull apart cleanly.
That said, what you can do is enhance parts of the sound that will make one voice more clear than the other. When we talk, the frequencies we use to distinguish words vary from person to person. If you use an EQ to reduce the frequency of the child's voice and boost the ones of the voice of the speaker you want to hear, you may be able to make out what they are saying more easily.
It isn't a guarantee though and if the voices are close, there likely won't be anything you can do other than have someone listen to it with well trained ears that might be able to make it out.
Basically what you want to do is:
- Find the range of frequencies for each person. The main problem is here since frequencies might overlap. You said that one of the speakers is heard through a phone, maybe it will result in more easily separable frequency ranges.
- Raise the volume of range of frequencies of the person you want to hear
- Lower the other range accordingly.
This is easily done using parametric equalizer which lets you choose any set of frequencies from audio signals, however I doubt there are any such software provided free. Also bear in mind that after all the voice still can be too low to hear\understand if there are too much overlapping frequencies.
I think your situation is that you have 2+ people on a recording, and they are speaking at different volumes.
I've come across three ways to do this (May 2015):
create 2 audio tracks (one for each speaker) and export to MP3, which apparently levels the volume (haven't tried);
Select certain snippets and amplify the audio for the quiet speaker (tutorial:)
Use Compression and Normalization tools in Audacity 2.06 (tutorial: http://www.jiscdigitalmedia.ac.uk/guide/audio-post-production-techniques-for-spoken-word#tips - scroll down to "Screencast of Compression and Normalisation using Audacity").
First I would try (3.) and then supplement with (2.) if necessary! No plug-ins required.
In the free Audacity Audio Editor, the free Nyquist plugin you can download is called "LevelSpeech.ny". It is designed to fix an audio interview where one person is talking with a quiet voice, and the other person is taking loudly, or yelling. It also seems to work where one person talks normally and then suddenly starts screeching and yelling for emphasis. In extreme cases where Clipping has also occurred, I run the "Clip Fix" plugin first. Then reduce the volume to around 80%. Then run "LevelSpeech" at around 95%. It's not perfect. But it tidys things up. Also, Nyquist plugins are usually interchangeable between audio editors.