Audio "Normalization"

[Robert Chandler]

I have an upcoming project where the client says audio must be "normalized to 0db."

I don't think that this refers to the audio scale, but rather to a global setting in software (I could be wrong).

We are using Edius 4.6 Broadcast. It is such a big-footprint software package that it MUST be able to do such a thing, but I'll be dogged if I can find it to see if I'm set for 0db, or if I'm headed for trouble.

Or is this a Procoder issue?

Any help would be GREATLY appreciated.

Robert Chandler
WAFB-TV News Operations
Baton Rouge LA
[email protected]

[Stereodesign]

Normalised to 0dB means that the audio peaks at 0db rather than at some other level, -12dB being another common level.
Any decent audio software should have a normalise function with the ability to choose the max level.

[cuervo]

It's deriguer that all digital audio is normalized to 0 dBfs. This is quite different than analog audio, which is a different dB scale. In digital audio, the PPM signal should never exceed 0 dB or hard clipping will occur. Most hi end DAW will include a normalize function. I don't believe Edius has one. Look for a plugin to accomplish this function, or invest in something like Soundforge, which has a very nice normalizer.

[LightPrism]

Edius' built in audio filters does not include a Normalize Filter.

But Edius does have the capability to accept many "VST" filters.
There are a many, many Normalize VST filters available on the internet - try a Google search and you will see this. Many are free and work fine within Edius ... some are problematic in Edius.

Since you appear to be working in the broadcast arena, I'd suggest that you purchase the new Blue Audio Filter pack that is built to plug into Edius.

[pjsssss]

Since you appear to be working in the broadcast arena, I'd suggest that you purchase the new Blue Audio Filter pack that is built to plug into Edius.

Let me add that Digital Juice has a special with NewBlue AudioFX bundle for $149 (which is supposed to be 50% off)

[Stereodesign]

remember that if all your audio files are Normalised to 0db then when you mix them on the timeline the overall mix will sound awful unless you adjust the volume of the files relative to each other so that the overall peak level of the finished audio is 0db, its a peak, not average.

[Robert Chandler]

Audio "Normalization"

I appreciate all the help...

For this initial test, we will peak audio normally (just reaching the yellow) and see how it opens on the receive end..

The human putting out the "parameters" info is more into sales than technical, and the information is suspect..

I find it interesting--given the wide sweep of the Edius software--that a seemingly standard feature like audio normalization is not part of the package..

Thanks to you all.
Robert Chandler
WAFB-TV News Operations
Baton Rouge LA
[email protected]

[biopic]

Let us not confuse "normalizing" with compressing. The only way to normalize an audio file is to scan it from start to finish and find the maximum amplitude. Then the volume of the whole file is increased such that the maximum amplitude becomes 0dB. This process cannot be done in real time, since you don't know what the maximum will be until you reach the end. Therefore normalizing cannot be accomplished with a VST plug-in - not even in Sound Forge, Wavelab or anything else - it has to be a separate 2-pass process.
That said, it would be really good if there was some sort of "normalize audio" tick box added to the print-to-file/print-to-tape options in Edius.

[cuervo]

biopic..

you've never, obviously, used soundforge to normalize an audio file.

[Rusty]

biopic is on the right track...
SoundForge is two pass.

I like to isolate the dialog and use Levelator

[Lemi]

Levelator it is. Since i found it (here, of course), I didn't even tried anything else by that matter.

[GrassValley_BH]

Rusty's right that biopic is right.

SoundForge and other audio applications can normalize because they have "knowledge" of the entire audio stream.

EDIUS's plug-ins, on the other hand, receive only a stream of audio frames, so they only know the past, but not the future.

Or to put the scenario another way, can you Normalize a live source?