Understanding how HarBal spectrum carving works

[jess]

Hi guys,

Just want to quickily understand how the spectrum carving works.

If I have say a 3 minute mix that has a good spectrum from Har-Bal point view and then I arbitrarily added a screaming vocals that lasted about maybe 2 seconds. The vocals peaks at 2-3KHz way off. Now, from the Har-Bal analysis point of view...would there be a sudden peak at 2-3KHz now?

The answer seems to be yes? If so, iIf I carve away (say I'm not even interested in even making the scream much audible) just the sudden peak will it destroy the rest of the proper mix?

This is rather hypothetical, but it helps me understand the mechanics and how I should or shouldn't carve certain aspect of the spectrum.

[HarBal]

If it's only 2 seconds out of three minutes then the peaking will only really show up in the peak spectrum trace. The average trace will not be affected by it much at all.

If you really have a big issue with just a single short section as in your example then your best approach is to split the track into sections and process each individually, otherwise you'll be penalising most of the track for the sake of a bad scream at the end.

This sort of track splitting is planned for the next major release of Har-Bal but that won't happen for quite a while.

Cheers,


Paavo.