Comments

Chienworks wrote on 10/11/2006, 5:16 PM
Would be nice. Kinda like in Sound Forge the graphic fade has settings for 100%, 200%, and 400%. If there was a switch like that for the gain in Vegas you could have it start lower in the event and drag it up.

Or, you can use the volume envelope.
ibliss wrote on 10/11/2006, 6:24 PM
Or normalize the event, then the only direction you need to go is down...
Cooldraft wrote on 10/11/2006, 8:25 PM
What I am looking for is an immediate visual cue like it gives you when you go down.
Cooldraft wrote on 10/11/2006, 8:29 PM
Normalize is good, but this waveform needs more in areas. Using a volume envelope is good, forgot all about that.
Chienworks wrote on 10/12/2006, 5:29 AM
Normalize doesn't help much when part of the event is already near 0dB.

The raw files from the sunday morning church service recordings i do have levels that vary all over the lot. Most of the sermon might be at -24dB so that i have headroom for occasional expostulations. Then there's the occasional fist-thump on the podium or page-turn right next to the mic. These little occasions will often hit -2dB. So, normalizing the whole event will result in those few moments at 0dB and the rest of the sermon still at -22dB. Doesn't really do a lot of good.

In less dramatic circumstances, during the praise & worship time most of the musical peaks hit around -6 to -3dB. However, between songs one of the team members may pray or do a reading, and these are usually about 15dB quieter than the music. Normalizing the whole event still leaves the speaking parts 15dB quieter. Going "down" from there isn't an optimal solution.

I use several methods to even it all out. We have a hardware compressor inline before the signal is digitized (and yes, those dynamics i mentioned are after the compressor). Wave hammer is amazingly good at handling huge volume variations in speech. Used less forcefully it's still very good for music too. In the end though, i split all the sections apart and process/tweak each individually. Loud peaks may be selected and process/graphic faded individually before dealing with the whole event. Especially loud bangs and thumps are often deleted completely so that they don't skew the rest of the processing.

It takes a lot of time and manual control to tame the volume swings properly.
Bill Ravens wrote on 10/12/2006, 7:14 AM
normalizing is worthless, except as the first step in post production. The real important number is the RMS volume. the only way to get the RMS volume up without clipping the peaks is to use compression and soft limiting. Any time vocals are involved, the dynamic range of recorded sound is pretty much more than a microphone and its support electronics can handle. It is rather surprising to me that camera AGC hasn't been replaced by a tunable compression filter.
Spot|DSE wrote on 10/12/2006, 7:33 AM
Normalizing in Vegas is somewhat worthless anyway as well, IMO. This is where Sound Forge shines, you can normalize on RMS.
Either way....I'm glad Vegas doesn't do what Cooldraft is asking for, it would easily contribute to truncated audio, and probably create far more problems than it would "cure."
Of course, the real answer is to work with your shooters and sound dogs to record proper levels in the first place.
Bill Ravens wrote on 10/12/2006, 7:48 AM
Spot..

hmmm...does normalizing in Sound Forge with the RMS option selected, clip the peaks or compress them? Seems like way back I tried that method of raising the general volume and got one hell of a lot of distortion and clipped peaks. The only tool I've found that works reliably is Ozones Loudness Maximizer, set to the right threshhold and with a margin slightly less than zero.
farss wrote on 10/12/2006, 2:35 PM
SF's RMS normaliser includes an option to compress if clipping occurs.