Audio Levels - Normalising Question

Organism Seven wrote on 8/29/2005, 6:29 AM
Hi,

I hope you can help me with this time consuming problem!
I often have footage where I want to apply normalise to the audio track, but because there is often a high peak caused by someone coughing, a baby crying out, chairs being moved etc etc, I have to zoom in and select either side of the high peak, split either side of it and then apply normalise so that it avoids trying to match the levels to the high peak glitch caused by the above type of noise.
Is there any way I can get this filter to ignore the high peaks which are nearly always 'glitchs' and just average out the sound across the whole timeline including all the edits/splits I will have made?
This will save me many hours of having to manually crawl along the timeline making all the adjustments that I am doing at present.

Thanks in advance for help you can provide.

Regards
Organism Seven

Comments

Spot|DSE wrote on 8/29/2005, 6:45 AM
Is there any way I can get this filter to ignore the high peaks which are nearly always 'glitchs' and just average out the sound across the whole timeline including all the edits/splits I will have made?

No. By its very nature, this would avoid what normalizing is supposed to do. You could compress it to avoid the coughs, with a fairly high threshold, and a very steep compression ratio, but that's about the only way to avoid the coughs. If you wished, you could also render the compressed track to a new track, and normalize that, I suppose.
Grazie wrote on 8/29/2005, 7:07 AM
What I do is make a "split" either side of the peak. I've also done some quick and dirty corrections 'cos I had too. It isn't THE correct way to do things, but I've been able to "pull" stuff back out of the "fire" this way .. . It's OK, Spot, isn't listen now - lol!

I forgot . .memory done . . Then go to the events properties and normalise there .. seeesh .. sorry lads!

Grazie

( edit . .senior thingy! )

Spot|DSE wrote on 8/29/2005, 7:45 AM
I don't understand, Grazie, why isn't your way the "correct" way? Organism's post is that he DOES'NT want to split segments out, which is what he's doing now.

As far as normalizing, I can't see very many reasons to do this on an event by event basis. This is why there is a free script (from Sony, I believe) to achieve normalizing on a selected or global basis.
Organism Seven wrote on 8/29/2005, 9:03 AM
Hi,

Where can I find this script?
Even if I find and try to use it, will it really achieve what I want?
I assume it will still just see the high peaks and use them as a 'guide' for normalising the audio.
The same as if I applied it manually?
So I would still have to edit on an event by event basis. Tedious!
You are right DSE, I am trying to avoid all the split segments just to normalise the audio.
Thanks.
Grazie wrote on 8/29/2005, 10:15 AM

Good point Spot! - I always thought that splitting was a messy way of doing this, and that normalising a whole track was the correct way. However, my way IS a way. Doing a "whole" track would be a monster.

Grazie

Chienworks wrote on 8/29/2005, 10:25 AM
I usually use Wave Hammer for this. If most of the track is around, say, -12dB and the few coughs and chair bangs are around -3dB, i'll set Wave Hammer for hard compression starting at around -11dB and auto gain compensate.
Grazie wrote on 8/29/2005, 10:29 AM
How different is Wave Hammer RMS to "Normalise" RMS within Sound Forge?

G
Chienworks wrote on 8/29/2005, 10:35 AM
Well, to be honest, i've rarely ever used RMS normalization. Wave Hammer seems to be a lot more intelligent about what it does. It seems to squish the sound less to achieve it's goal.
Grazie wrote on 8/29/2005, 10:37 AM
Thank you. I'll keep that in mind! - G