normal volume levels (RMS and peak)

[lukejs]

Hello. What would you say, as a mastering engineer, if you received a mix that had an average RMS of -25 , and a peak at -8 ? This was a soft ballad, and I was wondering if it was just recorded with levels that were too low.... Or is this normal to receive mixes with such low volumes, because of the high dynamic range of 24 bit recording? I´m thinking that the analoge gear should be pushed harder.... anyone give some advice ? Thanks.....

[MASSIVE Master]

That's not low - That's pretty normal.

Of course, you could bring it up a few and it'd be fine also. As long as it "naturally" doesn't clip (and preferably has a dB or two up top for luck) I can't think of a M.E. that wouldn't be happy.

[Ronny]

Originally Posted by lukejs

Hello. What would you say, as a mastering engineer, if you received a mix that had an average RMS of -25 , and a peak at -8 ? This was a soft ballad, and I was wondering if it was just recorded with levels that were too low.... Or is this normal to receive mixes with such low volumes, because of the high dynamic range of 24 bit recording? I´m thinking that the analoge gear should be pushed harder.... anyone give some advice ? Thanks.....

The problem with using RMS to relate perceived gain is that all songs will exhibit different results. A song that has a lot of dead air breaks where the singer is soloing a line sans instruments, a long fade in intro and a long fade out outro, can have a lower RMS and actually sound louder than one with higher RMS with more constant levels. Perceived gain is always relative to the level heard immediately before. If dynamics go from -20dB to -3dB in say 2 seconds, you will perceive a louder impact than if the dynamics go from -10dB to -0dB in the same 2 seconds of timeline. Although the peak is 3dB higher, the difference between actual dynamic fluctuation is greater going from -20 to -3dB. It's a 17dB difference where as the -10dB to -0dB is only a 10dB difference.

To use RMS more effectively you'd need to analyze many, many short sections and particularly at dynamic level changes in the song. People that analyze RMS over the whole song can be misled. Crest factor is much better at simulating perceived gain than RMS analysis. For example your -25dB RMS with peak at -8dB, is going to actually exhibit a crest factor that is closer to -17dB, for the simple reason that your -25dB RMS analyzer is taking into account the level between -8db and -0dB where you audio doesn't reside. It also takes into account no levels, for example if you include 2 seconds of dead air at the beginning and end of the song in the RMS analysis your RMS will be lower than if you leave the dead air out of the analysis and the music hasn't changed a bit. IOW, if you raised total gain from peak -8dB to peak -0dB and re-analyzed the total song for RMS, your RMS would raise in gain relative to the peak gain, your -25dB would than be -17dB RMS, but your crest factor would remain the same.