A question regarding how to deal with annoying peaks when normalising recorded wave files.
I record my band at rehearsals and gigs on my ZOOM H2. I record fairly low, in order to avoid distortion. Recording format is MP3 320 kbps
Most of the information is within - 20 dB on both sides of center as seen in attchm
When I normalize the recording, some annoying peaks are preventing me for getting the highest possible level. See here:
Untitled1 is normalised and the disturbing peak is highly visible.
In Untitled2 (the original recording) it can be seen that the original peak is around -16.6 dB and can be measured to last around 2 ms.
One way to handle this is to manually go through the file, select any disturbing peaks and select in menu: Process - Change level… to reduce to a suitable level. This is a tough job, when the recording lasts more than an hour.
Another way I would assume to work is to apply a Limiter effect, but this is where the problem starts. (Maybe due to lack of my understanding.)
When I select Process - Dynamics… I maintain the following settings:
Attack: 1 ms, Hold: 5 ms and Release 1 ms.
Threshold: -20 dB, Ratio: 10.0:1
With these settings I would assume that a peak at -16 dB would be processed and the rest, which is -20 dB or lower is left untouched.
But this is not what is happening. Actually the entire wave is processed and “squeezed” no matter how low the dB level.
When I play around with the Normalize before and after settings I can’t see the waveform is altered. The peaks are still there and no difference from just normalizing the original recording can be observed.
Do I operate the compressor/Limiter in a wrong way?
Is it simply better to manually reduce the gain of the disturbing peaks?