Hi to all

I've been asked to try to save a few days of recording (interview with boom and 1 radio mic): the radio mic is always (four days of recording, I couldn't believe it!!) saturated/distorted; to me, it seems like if the production mixer/boom operator had set up the receiver to line level but connected it to a microphone level input in the recorder.

Every time the interviewer speaks, the sound is over compressed, saturated, distorted (but it's not clipped, it's an analog problem not digital).

How can I improve a little bit the recordings and reduce the distortion? I already tried with izotope RX, with some multiband expansion, some eq... Any better ideas/workflow to try?

If I cannot save it with my tools, I advised the producer to go in another studio that have the big Cedar machine (the Cedar Cambrige): anybody has any experience with the cedar cambrige (I used only the DNS) that's willing to share and know if it could save those saturated recording?

By the way, the boom is not always an option, because lots of the interview where made in noisy places (market, on bus) and the signal to noise ratio is very poor.

Thank you in advance


4 Answers 4


Most Declipper plugs won't work in this case, because they're designed to work with the very straight, minimal harmonics of digital clipping. In this case you've got analog distortion which isn't quite the same. One thing you CAN try is using a De-Crackler plug-in. I've found the Waves X-Crackle the best for this, but the WaveArts works as well, and presumably the RX one (I haven't had an opportunity to try it yet). Try setting the values to maximum and see what happens. Follow that with some negative compression (i.e 1:2, 1:4 ratios )- I like the Waves R-Comp for this- and you can often minimize it, albeit not remove it...You can also try doing the negative compression first, sometimes that serves better.

  • X-crackle and the powercore de-crackle are the first things I tried.. And I had a bit of success... Negative compression is a nice trick, and sometimes worked, thanks. In the end we decided to take everything to a studio with the Cedar Cambridge, hope that they can save those recordings: they are the worst recording (made by a pro) I ever listened to... Commented Nov 29, 2010 at 15:52

Unfortunately when a sound is recorded and distorted as a result of excessive levels the audio information is simply not there to retrieve.

De-clipper tools make an educated guess as to what the peaks should have looked like, and rebuild them based on what audio information is available i.e. the waveform just prior to distortion. Apparently they are only really effective when there is light clipping/saturation, and you already mentioned that you have used Izotope RX so I'm guessing that the saturation is pretty bad.

I think CEDAR is your best bet without re-recording the whole thing. Never used it myself but apparently a lot of the noise removal features are modular, or in other words they are add-ons to the basic package. If you do go down this route make sure the studio you use has the de-clip module. From what I have heard/read CEDAR is in a league of it's own, so if that doesn't fix it then nothing will.

Sorry not to be of much help, I pretty much re-iterated what you have already said but hope it works out for you!

EDIT: Here is a short movie showing how the de-clip module works. Unfortunately no comparison of audio though:



Obviously for the benefit of others with the same problem, since this was 18 months ago... (Hope you managed to make a difference, by the way!)

The Cedar is really amazing and might help.

If it's at all a harsh-sounding distortion and you're there with a Cedar anyways, you might try de-crackling and see what happens.

Using an expander might do something for you but I suspect not.

Used in combination with another successful approach, you might help the quality at least a little with some eq, because whatever frequencies are emphasized by the mic and pre-amp, when gained up too far, tend to splatter all over and really dominate when squashed. You may also find that the crackle of the distortion lives predominantly in one frequency range and you can knock it down a little with eq. (Or try the C4 in this capacity.)

Desperation move: try blending a little of the lav mic in with the boom for clarity. You can experiment... maybe band-pass the lav's signal severely so you're left with mostly consonant sounds and sibilance, and see if you can blend this in to help make the boom audio more intelligible in the quieter spots.


You can try to fix it by editing the volume automation. Where you feel the signal is too much compressed, reduce the level. It is a painful process. I once made it for a 100 min documetary Maybe you can use RX as a guide. You schould see gain manipulations outside the human voices spectrum in the spectrogram.

Good luck...

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.