Advice on fixing vox tracks?

[mdurrie]

I just joined today, hoping someone has some words of wisdom on this problem. I spent a day recently recording a number of tracks with a female vocalist. I was a bit rushed, and to make a long story short, didn't notice until today that she was overdriving my Apex condenser mic during the loud passages. There's no clipping of the waveform itself, but there's a definite crackle every time she really lets fly with the amplitude. Obviously, what I SHOULD do is get her back in my studio, move the mic about another 12 to 18 inches away from her (or maybe three feet -- she's LOUD!), and do it over. However, this is impossible, since she flew back home (about 1,000 miles away) yesterday. I got some really good performances down, so I want to try salvaging what I can. My first instinct is to try finding the exact frequency of the crackle and EQ it out. Is that a good plan of attack? Other ideas, tips, scorn, sarcasm, etc.? Thanks!

[Chibi Nappa]

I'm afraid you're out of luck... Those crackles can't be eq'ed out.

[Hi_Flyer]

there are some de-clicker plug-ins that might work... I know that Wavelab has a click remover, I think Soundforge does too. They might leave small artifacts in the audio though.

[Ford Van]

Originally Posted by mdurrie

I just joined today, hoping someone has some words of wisdom on this problem. I spent a day recently recording a number of tracks with a female vocalist. I was a bit rushed, and to make a long story short, didn't notice until today that she was overdriving my Apex condenser mic during the loud passages. There's no clipping of the waveform itself, but there's a definite crackle every time she really lets fly with the amplitude. Obviously, what I SHOULD do is get her back in my studio, move the mic about another 12 to 18 inches away from her (or maybe three feet -- she's LOUD!), and do it over. However, this is impossible, since she flew back home (about 1,000 miles away) yesterday. I got some really good performances down, so I want to try salvaging what I can. My first instinct is to try finding the exact frequency of the crackle and EQ it out. Is that a good plan of attack? Other ideas, tips, scorn, sarcasm, etc.? Thanks!

Crackles usually mean clipping SOMEWHERE! So, maybe you didn't clip the A/D converter, but certainly you clipped the preamp at some point.

You might be able to fix it a little bit, but it will be VERY slow going!

In applications like Wavelab, you can redraw the amplitude of the waveform at the sample level. Put another way, you can actually raise/lower the volume of the audio at the sample level!

Most likely, you will find that the spots where you hear the crackle that the waveform on the screen is a flat line. If you were to adjust the volume, per sample, of those spots to be lower in volume, or maybe just not have so many samples in a row that are flatlined, you MIGHT improve things a bit.

This actually works very well if you had clipped the A/D converter. When you clip the a/d converter, the waveform will be at full volume for at least 8 samples in a row. Simply by adjusting the volume of those samples so that they go up to full volume in a "spike" you will remove much of the "audible" distortion associated with the clipping. No, it does not always TOTALLY get rid of it, but you would be surprised how much you can fix! I have had to restore some audio tracks that were clipping at the a/d this way. I would have much rather recorded the tracks again, but that wasn't an option.

[mdurrie]

Originally Posted by Chibi Nappa

I'm afraid you're out of luck... Those crackles can't be eq'ed out.

Is that because they'll be at the same frequency as part of the vocal? It sounds to me like a simple de-essing might remedy the problem to a large extent, but I haven't tried anything at all yet.....

[mdurrie]

Originally Posted by Ford Van

....Most likely, you will find that the spots where you hear the crackle that the waveform on the screen is a flat line. If you were to adjust the volume, per sample, of those spots to be lower in volume, or maybe just not have so many samples in a row that are flatlined, you MIGHT improve things a bit. ...

Hmmm, interesting. So, even if it's not clipped in the sense that the waveform is being visually cut off, I should still look for loud "flat spots," indicating that clipping has occurred somewhere in the signal path?

[mdurrie]

Originally Posted by mdurrie

...I should still look for loud "flat spots"...

OK, I went in and examined the waveforms. There is no visual evidence of any sort of clipping, no flat spots. Just that crackle in the louder passages. I believe the problem is sonic (i.e. originates in the mic diaphragm or something like that) rather than electronic. I assume due to the lack of any other suggestions that no one has any other thoughts or experience with this. I'm going to try some "surgical" EQ....

[Ford Van]

How far did you zoom in?

Basically, the audio is going to have to have many samples in a row that are the same volume. It IS there if you zoom in. You are looking for stuff where at least 8 samples in a row (to see actual samples, you will have to zoom nearly all the way in on the audio) are the same volume, and possibly the loudest volume.

[gullfo]

another thing to consider - any takes in the song on the same lines where you didn't get crackles? if so, you can clone those sections and put them in place of the crackling ones... otherwise, as pointed out by several folks, you're going to have to play with it on a nearly sample-by-sample basis to get them partially cleaned up (i say partially because there will be some artifact left regardless - either loss of frequencies or a muted crackle) and in the context of the mix might not matter...

[grn]

yeah... copy and paste