I recorded someone talking on the subway but the voice is lower than the subway noise. Can I reduce the subway noise and increase the voice? I tried the noise reduction technique in effects but didn’t really work that well. Thank you.
Audacity can’t split a mixed performance into individual voices, instruments or sounds.
Noise Reduction fails if both sounds are changing. You might be able to find a tiny section of the recording where the train speed didn’t change and you didn’t move the recorder. Noise Reduction works by getting a clean sample of the noise (Profile—without voice) and then apply correction to the whole show. That might give one or two seconds.
If this is the New York subway, there’s never a stretch of track where the train speeds don’t change.
OK, maybe Rockaway, but that’s it.
You picked a terrible studio for a recording.
Also there’s the side issue that Audacity can’t be used for surveillance, law enforcement, or conflict resolution.
Koz
Oh, I was recording my friend talking.
What program should I use to split the voice and the background?
What program should I use to split the voice and the background?
This isn’t a help desk. It’s a forum. Users helping each other.
We can wait for someone to arrive with the solution, or you can find one and post back how you did it. There will be many thousands of people wanting to know the process.
Koz
Was this a video? You didn’t say. If it was you can use a Hollywood trick. They have a thing called “Looping.” where the performer sits in a quiet room and watches segments of the video again and again (looping) and announces clear perfect replacement words matching up the lips. You take the perfect words later in your video editor and put them in as the New Foreground sound with subway noises in the background.
Koz
Thanks but this isn’t what I’m trying to do. I want to keep some of the background noise, I don’t want to redo it. Thanks everyone for input, much appreciated.
If you absolutely have to do this kind of show in real time, there are two microphone tricks. NPR discovered a novel use of long-distance shotgun mics. You use them short distance. This is good for pulling a good interview out of a very noisy environment.
The other is a lavalier microphone. That one is worn on the chest and it works by being much closer to the voice than other microphones.
You can get lavaliers in directional form and they’re even better than regular ones to split the performer from background noise. But they’re also little time bombs because if you shift your jacket or shirt a little, you can lose the voice.
Neither of those is particularly cheap. The shotgun engineer is holding about a grand-usd in his hand not counting the recorder.
This is where you tell us you can’t reshoot it and this is where I say you have no show. And don’t record that way again.
We can wait if someone else wants to take a crack at it. Can you post a sound sample?
https://forum.audacityteam.org/t/how-to-post-an-audio-sample/29851/1
Koz
Thanks but I will just use what I have, background noise and all.
Audacity is pretty good at reducing low level noise, but if the noise is louder than the voice, then I doubt that any program can make a significant improvement.