Processing live audio podcast


I’ve recorded a spoken word and music event live on a Zoom H2 recorder. The recording is ok but the levels are problematic. I.e. the applause is really loud compared to the performances. The recorder was not held right next to the performers so as not to be intrusive.

I was hoping someone might be able to help with some suggestions about how to process this to get the best results. We were hoping to use an mp3 recording so people could listen back to the evening’s performance and maybe even use a podcast for others who are interested.

I am pretty new to this and messing around with compresser/hard limiter and amplify but not really sure what i’m doing. would really appreciate some guidance on best way to process this in audacity. I’ve attached a short clip of what I’m working with.

You could cheat. This is what happens when I drag-selected the guitar only and applied Effect > Amplify > OK.

I assume you want something more automatic. I’d try Chris Compressor.

Chris’s Compressor

The recorder was not held right next to the performers so as not to be intrusive.

You might need to be intrusive. Live performances are always a pain because “theatrical” and “recording” are different. There are some interesting things you can do with the microphone/device to stay out of the way. Everybody sees the big boom and mic stand with blast filter and cables hanging all over the place, but you can fake a stage microphone on the floor.

Nobody pays any attention to the sound people until it’s too late.


  1. Replace the (stereo) applause with a mono-ized version, this gets rid of the intrusive stereo effect which occurs on the distorted bits of applause.

  2. Cut out the most obvious bits of distorted (now mono) applause.

  3. Use Audacity envelope control to make the guitar louder and the applause quieter.

  4. apply a hint of chorus and stereo reverb, (adds pseudo-stereo effect to mono applause)

As the clapping is significantly louder than the performance, the AutoDuck effect can be used to drop the volume of the applause.

  1. Duplicate the track
  2. Select the upper track and apply AutoDuck with the threshold set at -12 dB and the Duck Amount set to about -20 dB. Leave the other settings at default.
    You could set the Duck amount even lower to make the audience quieter, or apply it more than once with the the “Inner fade down length” set to 3 seconds on the second pass.

The advantage of this technique is that the sound of the actual performance is not affected at all.

In this example, the click at the end of the last guitar note was reduced manually with the Envelope tool as that click cannot be removed automatically.

adding a 100Hz high-pass filter on the applause helps mitigate the distortion …

Many thanks for the advice. Looking forward to giving these processing ideas a try later on (at work at the moment)

Am confident this can sort the contrasting levels.

Thanks for the idea about addiding additional chorus/reverb. I was wondering if anyone thinks any additional compression or other fx need to be added on the guitar? I also have sections like this where it is speech which is similarly v quiet too. Would they require different processes?

(Also, need to look at that idea for stealthily getting the mic closer to the action next time!)

One of the problems with the Zoom series is you can’t be there to set levels unless you’re hand-holding it and even then. My first pass clip illustrates one of the problems with that. The applause is the “right” level although it has overload points in it (did you do that, or did you let the Zoom auto-set?), but the music is seriously low. My simple volume boost brought up noises not immediately apparent in the original.

This is very similar to a lecture recorded from the back of the hall – maybe minus the echoes. The complaint is always the audience noises are far louder than the show.

Another complain of the Zoom is that the external microphone amplifiers inside it aren’t that great.


Hi Koz. re: the applause level, the Zoom was actually closer to the performer than the audience - but it was a small room. I had the recording level at 100 as I was trying to avoid the distorted clipping. (not full) I had the Auto Gain Control SPEECH setting on, (most of the event was spoken word, with a couple of acoustic musical perfomers) and the Lo PAss Filter.