My name is Jan, I live in the Netherlands and I am a newbee in audio editing. Recently, I recorded a choir with music using the automatic recording setting on my recording device. As a result, the audio is fluctuating in volume (pumping effect?). Is there a way to correct this?
It is very difficult to fix that problem. You may be able to make some improvement using the “Envelope Tool”, but it won’t be perfect.
Theoretically an expander can automatically reverse the pumping-compression of “Auto recording level” (a/k/a AGC),
but in-practice manually correcting it with envelope-tool is quicker …
Thanks for your help, I will try your suggestions.
Theoretically an expander can automatically reverse the pumping-compression
In the real world you never know the exact compression characteristics so you’ll never get the complementary settings right and you’d probably make it worse!
And, you can’t fix the clipping/distortion that happens when the sound suddenly goes from quiet-to-loud with the recording volume cranked-up, but manual adjustment/correction should sill help.
…There are systems that use complementary compression & expansion (or at least there were in the “analog days”) but they are carefully designed to be complementary with known and controlled parameters.
Recently, I recorded a choir with music using the automatic recording setting on my recording device.
For the future… If you’re recording something critical and there’s no possibility of “take 2”, I always recommend running a back-up recording system in parallel. That’s especially true with computers because computers are the least reliable things we own and there are lots of settings & “variables” and once in awhile things “go wrong”. And with a computer, sometimes you don’t know anything was wrong 'till the next day. (However, Audacity doesn’t apply effects in real-time so you wouldn’t get that AVC problem. )
To create my own parametric dynamic expander I remembered old, from hifi Compact Cassette decks well known, Dolby noise reduction systems.
They are working as follows:
During music recording the Dynamic of the Original is compressed by an AGC by about -10dB. (simplest Dolby B do this only within the especially noisy frequency range over about 500-1000Hz).
During playing the record the mirror process is done as the here discussed dynamic Range Expander to get back the original signal with now -10dB reduced tape noise level. This happens quasi by a subtraction of the compressed signal from the AGC-restored signal.
This Process I Do in Audacity as follows:
- Open or import the original File,
- normalize the track to 0dB,
- dublicate now this track to a second one,
- invert the second track,
- now compress the second track with a ratio of 3:1 (10dB),
- very important for well-doing: normalize now the second track to -20dB,
- mix and render both tracks to a new one,
- normalize now this new track to 0dB - ready!
It is important for not to be heared this harsh intervention to parameter the compressor vary carefully.
I used an attack time of 0,1s and a fade away time of 2,5s. These values will very seldom to be heared because they correspond to the accoustic properties of most concert halls. shorter time constants will non linearly distort the signal more.
Higer Expansion than 10dB (3:1) is possible, but also will be heared as “pumping” and singal distortion in extreme case.
The hole process I wrote to a makro to automate it. It works very fine in pop music as well as classic like organ or piano music. I am very satisfied by the good results.