I have a digital recorder which can simultaneously record into two files: one with normal volume, and another with -12dB volume (“backup recording”). I was wondering if it’s possible to smoothly blend the best bits of these two tracks, and what would be the easiest way to do this, in Audacity. Basically, I would like to use the silent track for the parts where the louder track has clipped the signal.
In photo editing, for those that have experience in it, this operation would be equivalent to exposure blending; replacing blown areas in an overexposed image with those from a darker exposure. But perhaps in audio processing this is a much more difficult operation?
What you are describing is called “crossfading”, which is basically fading one track out while another track fades in. This can be as simple or complicated as you like I’ve started writing an article about crossfading here: http://wiki.audacityteam.org/wiki/User:Stevethefiddle/Fade_and_Crossfade
I think the envelope tool would work too, but it can get rather laborous if there’s a large number of volume spikes in the recording.
I’m interested in a more automated way of doing this. I’ve been trying different values in the “Auto duck” filter, but I’m thinking the single, hard threshold value might be a problem; the effect would need to be more gradual to not sound too noticeable. That is, the filter now waits for one threshold, and then applies the complete ducking effect, while I think it would sound better do this continuously but gradually, fading out the track by an amount relative to its current volume.
Yes, the envelope tool will work (though laborious as you say), provided that the recording level does not go so high as to cause distortion.
Similarly, as long as the audio is not distorted, you can partially automate the process by using a “dynamic compressor” effect.
See: http://manual.audacityteam.org/o/man/compressor.html
Yes, I suppose applying the compressor filter on the backup recording is one way to achive a mostly similar result.
Here’s an illustrative screenshot related to the original question. The upper stereo track is recorded at normal volume, and the lower stereo track is the backup recorded at -12dB.
Wow, someone is LOUD
Even the -12 dB track is clipped in places. For future reference I’d suggest setting both versions 12 dB lower than they are now. You may then be able to get a usable recording from just the (then not clipped) quieter track.
How worthwhile it is for you to two versions depends on how you are recording them. You may be as well to just use the one quieter track.
I’ve started writing an article about crossfading here: 
