I am just learning.
I have amplified the sound on my cellphone recordings but now I have a noise problem - car noise, traffic noise, they were made on the road.
I do a noise reduction and it is good but more needs to be done.
I select some more noise and extract it and find it is too much - the remaining sound is hollow, bad.
But there is still a thread of pure noise. A slender thread now because that previous reduction took away a lot.
I am thinking I want to take that little thread out. Rewind that last extraction and this time extract only this ‘thin thread’.
But how to do that?
I tried to copy the thread, duplicate it, but then when I went to choices I found I couldn’t undo the noise reduction without first undoing the duplicate - so I lose the selection that I was going to use for the noise reduction thing.
I’m assuming I’d be able to do that? Use a clip on a second track to give to the noise reduction function in order to reduce noise on a different track?
Well that was/is my plan.
If it can be done how to do it? Will I have to save that clip to disk and then undo the duplicate and then pick it up off the disk and use it for noise reduction? Is that the way to go?
Or there’s a whole other technique I ought to know about?
Noise reduction is intended to reduce constant, low level noise, such as tape hiss.
Noise that louder and/or varies over time (such as traffic noise) must be avoided because it cannot be removed by software without causing noticeable damage to the remaining sound. You need to re-record somewhere that there isn’t traffic noise, or use a directional microphone close to the thing that you are trying to record so that the relative noise level is acceptably low.
If the performance is relatively even in volume, you might try Noise Gate.
Noise Gate tries to recognize low volume noise as “evil” and then tries to remove it. This can work OK in a relatively well-behaved performance, but if the background volume is bouncing up and down and the performer is inconsistent, it stops working.
Field Recording is a special specialty.
The person on the left is holding $1000 usd in his hand.
That’s a shotgun microphone with built-in recorder (on the bottom) and “dead cat” wind screen.
Matt (on the left) is a broadcast technician. The joke is every time Tom (on the right) goes out without Matt, his shows fall apart about half the time. Matt has a good deal higher success rate.
Oh. Right. The original question.
When you get an original raw recording of a performance, make a copy of it and move that copy to Some Other Place. Thumb Drives, External Hard Drives, Cloud Storage all work. It’s a New User mistake to shoot something, apply corrections and then post it all in one swoop. If Audacity freezes or crashes anywhere in the process, you can have zero show. Get the microphones, go out, and shoot it again.
Note it’s good to Export a WAV sound file for this backup job. Audacity Projects are not recommended. Don’t put most punctuation marks in your filenames. Today is 2019-05-12. -Dash- and Underscore are OK.
As you get the swing of this, open the back up copy once or twice and make sure you can. Nothing is more fun than accidentally creating a trash backup.
Thanks a lot for all of that. Very interesting and informative.
I’ve learned that I can indeed import an audio, use it for a noise reduction ‘template’ or whatever it’s called and remove that pattern from a completely different track. I did it. It’s fine.
There’s this one little ‘sub feature’ or something is what I’d like to know about. I didn’t explain it well before. I’ll try again in different words:
If I do a noise reduction to a track I finish with a ‘simpler’ track. My words. There will be sections that now are almost empty of sound shall we say. Nearly ‘empty’ or ‘quiet’.
I can do a further noise reduction on this ‘simpler’ track by selecting one of those ‘quiet’ sections and finish up with a third and final track that will have all those ‘quiet’ sections removed or not removed but totally silenced.
Now here’s what I want to do. What I’m talking about. I want to do that second noise reduction from the original. Missing out the first.
What’s the easiest way? Way with the fewest keystrokes?