Reducing Audio Noise

[Donald Harvey]

I have recently shot a short film and had to use my shotgun mic off of my Canon XH-A1, i hate doing it but this time i had to because the place i rented from gave me a faulty mic and XLR cable...screwed me.

Anyway, now I am in the post stage and have been listening to my audio, of course it is really low, and when I jack up the audio through right clicking and pulling up the audio gain or pressing the normalize button, I have this horrible back ground hiss or noise.

Can someone give me or teach me or give me a link on how to effectivly reduce that?

Thanks

Harvey

 

[Deacon]

I found this link on youtube and the tutorial deals with submixing. Don't know if it'll help, but just thought I would send it along. http://www.youtube.com/watch?v=YBfPMrF9Uig

Deacon

 

[j1clark@ucsd.edu]

Anyway, now I am in the post stage and have been listening to my audio, of course it is really low, and when I jack up the audio through right clicking and pulling up the audio gain or pressing the normalize button, I have this horrible back ground hiss or noise.

I'm not in to audio heavily, but I've done a few things to extract out sound from 'background noise'. The problem in your case is that you have a 'low signal to noise' ratio, so 'pulling up the audio gain', will not only increase the 'signal' but also the 'noise'...

So, what you need to do is to use a sound processing tool to attempt to change the relationship between the 'signa' and the 'noise'.

One filter is called 'noise gate', which is a bit hard to handle for some people control.

I've not used the sound tool that came with my CS3 Production Bundle, but the sound tools that I've seen elsewhere don't seem to come with a spectrum analysis widget. In addition they really don't come with a 'sonogram' widget. A sonogram is not something one looks at developing fetus or the like... it is a representation of the various frequency components of an audio track, and presents them in both 'frequency and time'.

Hence one can easily see where a 'voice' print region of the audio track is, and where there is just 'background', or even obtrusive sounds of random lenght.

With that sort of representation one can design a set of audio filters to extract out what one is looking for, or at least try to make the best of a nearly impossible situation.

But back to noise gate. By setting the various thresholds, one can clip the sound in such a way that the 'hiss' drops out from between the gaps of the person speaking, and it may 'clean' it up enough so that by rasing the amplitude of the clip voice utterances, the track is usable.

Now, in addition to that... after the noise gate is setup and 'the best that can be done is gotten'... add back a bit of noise for those 'gaps' in speach so the track doesn't sound clipped, but obvioulsy the 'new noise', is low enough that it does not interfer with the voice, but gives the listener the impression that the audio track was not heavily processed.