|
Short answer: Don't do either of them.
Long answer: Generally, you never want to use a straight normalisation process on a final mix at all. Instead, use a compressor followed by a limiter, as has already been mentioned. The only time I use a normalisation process is to quickly and temporarily raise the volume of a quiet sound so that it more closely matches other, louder sounds.
Secondly, don't use a RMS normalisation process either. As has also been said already, this is just a poor man's maximiser, indicating that you're much better off just using a dedicated limiter.
Just for the extra information though, if you did normalise using the average RMS level, generally don't set the value higher (i.e. closer to 0 dB) than around -12 dB. A value of around -14 dB can be considered nominal, though you can usually get away with a final RMS of around -12 dB at the loudest (the closer to 0 dB, the louder the volume becomes), though this does depend on the overall makeup of your track. If you have a really loud section for 30 seconds, followed by 3 minutes of low volume material, then picking an average RMS of -12 dB is obviously not going to work very well. In that case, you could attempt to normalise to an RMS value closer to 0 dB, though take care not to degrade the loud section too much.
Probably a better way of checking the average volume is set optimally is by using a VU meter. Enable the VU meter in Sound Forge and set the scale to Extended VU. Then go Options-Preferences->Other, and select -14 (EBU) in the 0 VU (+4 dBU) level box. Now, when playing your material through, ensure the loudest portion of your track peaks at around +4 dB. If you find it still isn't quite as loud as you like, type -12 where you entered -14 before and aim for a maximum of around +4 dB on the VU meter.
___________________
Put an end to the loudness war. Don't limit or compress your mixdown until mastering; leave the master channel alone.
Contact me for mastering.
|
