Citaat:

Thanks abaudio. You've explained this very well.

Am I correct in saying that peak-normalizing can be compared to simply turning up the overall volume of a sound system? That is, everything gets lifted by the same amount but the relative differences between individual sounds remains the same. For example if I have two sounds, one and 50 dB and one at 20 Db, and then turned up the volume so that the 20dB one is now 40 dB, then the 50 dB sound would now be 70 dB (the difference remains at 30 db independent of the volume). From a physics point of view this makes perfect sense to me since +3 dB represents a doubling in sound intensity. Thus when a sound is increased from, say, 70 dB to 73 dB, the intensity has doubled.

If my above interpretation is correct, this means that peak-normalizing an audio file just brings it into a workable audio region. The relative difference between individual decibels is kept constant. Thus, any noise present is not made "worse" by comparison to the surrounding individual sounds.



Well if you would have one sound panned hard left and one hard right you could maybe say that, but of course if we speak of a mix of two sounds being equally boost it is the mix that gets a boost, not the individual sounds. Either you normalize a mix (of several sounds) or a single track. The idea is that you maintain the dynamics. In other words, if the dynamic range (lowest level to loudest level) of a track is 15dB , it still is 15dB after normalizing. So the peak as well as up to the smallest level got the same boost. If you say the intensity gets a boost, I would relate that to the power within a track (sorry, maybe I misunderstand that, English is still not my first language), but if that is what you mean, then you would boost the RMS. Peak normalizing relates to overall volume, while RMS normalizing relates to (average) power in the track.

Citaat:


Disagree. You always should consider the side effects of whatever action you're taking to treat a sonic problem. This is especially going to be a big fat problem with recorded real acoustic instruments. All kinds of noise, squeaks and stuff are going to get turned up as well. Then, you have a new problem on your hand, no matter how impressive your meter level looks afterwards. Of course, those noises aren't added to the track by Normalizing it, but now they're audible and more pronounced.Therefor it "could" sound worse very much so depending on the quality of the original recording. If there's a need to boost track that much, you might as well re-record the track if it's an option.




This completely makes no sense. If you have to move up your slider to make the recording audible, you move up those noises just as well with it. In that way fader adjustments have the same effect as normalizing. In case that the "soft" accoustic recording can stand out enough without normalizing then a normalized track can be corrected by pulling down the fader and with it the noises. The results stay, no matter what, the same. The only thing that makes sense in your statement is "re-recording a track" and that is exactly what I said before. And most of the times that will be the case if you have to boost a track alot. A track with too much noise is simply a badly recorded track. No matter what those noises are. Noise as in hiss is a recording on an inputlevel that was too low, or e.g. squeeks that are too loud could be prevented or minimized by a better positioning of the mic or just by using a different mic. As I said before, peak normalizing does nothing more, nothing less than "putting up the fader", equally for all frequencies, sound level, dynamics. Also note that all the time I talk about peak normalizing, not RMS...

I'll be back...