Help with mixing the best from two sources


I have only recently discovered Audacity and find it great, I am just getting used to moving around the software and trying different things to see what they can do.

I have a project where I have two audio sources, a Canon 5D MkII with Sennheiser MKE400 mic and a Zoom H4. Unfortunately the sound from either is not good due to poor recording set-up and I want to take the best elements from each. The audio is of a live blues band playing loud in a small venue.

The H4 is on top, no clipping but no dynamics either. The 5D is on the bottom, loads of clipping but no bass.

I have experimented with Normalise, Leveller and Equaliser and have improved the sound (or at least changed it - not entirely sure it is better!). I am not sure which order I should apply the effects or it there are any others I should use. I would be grateful for any suggestions.

I am not expecting a studio quality recording from these two poorly recorded tracks but I would like to make the best of a bad lot. Also there is a slight sync issues the H4 tracks is .2 seconds out over the full length but I can fix that ok.

This is my first post here, so apologies for landing with a question on day one. Thanks in advance for any suggestions.


The H4 is on top, no clipping but no dynamics either. The 5D is on the bottom, loads of clipping but no bass.

Mixing the two is not likely to help anything. You’d probably get phase issues (comb filtering). And the comb filtering will change/drift with the time-drift (like a slow flanger effect).

It looks like the H4 waverform is clipped too. It actually looks more clipped than the 5D waveform. It’s just just clipped at a lower level (below 0dB) for some reason… Maybe the analog signal was clipped. (The only way you can “accidently” reduce dynamics is with clipping.)

I have experimented with Normalise, Leveller and Equalizer

The Equalizer is likely to have the biggest effect and be the most useful. You can boost bass with the Equalizer.

Audacity has a ClipFix “effect”, but since we can’t know the height or shape of the unclipped wave-tops, you can’t really undo clipping. But, it’s worth a try.

Normalizing (or Amplifying and setting peaks to 0dB) doesn’t change the “character” of the sound, it only adjusts the volume. It never hurts to normalize, especially after EQ or other effects that might increase the level above 0dB and cause clipping. (In order to prevent clipping, you have to reduce the volume before it’s saved/rendered.)

The Leveler is usually not a good effect for music. Music is supposed to have loud parts & quiet parts (dynamics), and “automatic volume control” can often really mess things up.

Some (dynamic) compression might help. It works something like the Leveler, but faster. It’s usually used to make the music “louder” or “more intense” by boosting the average/RMS level without boosting or clipping the peaks. Most recorded music is much-more compressed than live music. (And, the lack of compression is one of the things that makes live music sound better.) Modern popular recorded music is highly compressed, and radio stations add more commpression.

If you want to try, you can use high-pass & low pass filters to create one file with bass-only and another with no bass. And, the mix them. But, I wouldn’t expect this to work very well either. Filters introduce phase issues of their own.

Thanks for the reply DVDdoug.

You are right about the H4 clipping below 0db and I don’t know why, I have only used it a few times and never had the problem before I must have changed some setting or other, Back to the manual on that issue I think.

Unfortunately that performance was the ridiculously loud, even the musicians were complaining, so I don’t think it was ever going to be marvellous, I will try your suggestions for Equaliser and see how that comes out.

Very interesting info on Normalise, Leveller and compression. I am still playing around with Audacity and that was very helpful.



I’m not sure about the H4, but on the H2 there are two recording level controls. There is a High/Medium/Low switch that sets the microphone gain (this is the real input level from the microphone), and then there is a recording level control that works from the buttons on the front. The latter control simply scales the signal after it has been converted to digital. If the first control (the switch) is set too sensitive, the the audio will be clipped before it gets to the Analogue/Digital converter, and regardless of the recording level setting with the other buttons this clipping will not be corrected.

As I say, I’m not sure if the H4 works the same way, but it would be worth checking in the manual for future reference.

Thanks Steve, I downloaded the manual, and will review later.

I know I had the manual switch set to Lo and the levels from the menu set to 50 (down from 100). Reviewing the signal strengths it looked like the sound was peaking below max, therefore I assumed the audio would be ok - my mistake.

Thanks for the info.


Ear Bleeding Loud performances are a very special capture case. Even with the volume controls and switches all the way down, sooner or later you’re going to overload the actual microphone capsule giving you crunchy sound and clipping even without the red bands in Audacity.

Ribbons microphones just break unless they have some restriction to prevent it, Condensers (like the electret and probably the microphones in the Zooms) have two tiny moving plates and the plates touch if it gets loud enough. Dynamic microphones like the SM-58 do well. They start internally stressing and compressing their movement usually with minimal sound damage. Once when we were bored, we found that we could scream into an EV-635A…

…and it could be heard at the other end of a broadcast loop (radio telephone cable) without the microphone amplifier.