Which RMS levels to concentrate on during mastering? Total or Average?

[DorphonDalfir]

I

[Lee Knight]

Average is what I defer to. Total, while informative, could be pretty misleading to use as a guide to the perceived volume of a track. It looks like you've analyzed songs from your collection. I find that to be very informative as to what works for me and more importantly, what doesn't. One danger in using figures like this from an analysis app... you can stop actually listening. It's really a slippery slope.

 

Make sure to constantly A/B with stuff you like, while closing your eyes and only using your ears. I got pulled into that trap for a while. But using those numbers is a great way to really start correlating hard info with your perceived notion of volume.

[alphajerk]

you are reading numbers too much. the average over what track? does it have dynamics? if so, how long does it very between loud and soft? that will affect the average over the duration.

 

i pick the loudest section, get the compression/limiting nice and steady on that part and then the softer sections still have room to breathe... then again, i do this while mixing and dont bother with leaving ANYTHING to mastering if i can help it.

 

you really have to listen however. a sparse song you can literally SLAM and will be WAY louder than a really dense song. the really dense song will become "soft".

 

just do some A/B to tracks similar to yours that you are looking to match. play it by ear and see what you can get away with and quit looking at numbers. numbers are useless.

[Lee Knight]

 

numbers are useless.

 

 

Or... can be really helpful if you use them to inform you. But I agree with you. It's got to be the ears. Those numbers can really help educate you as to quantifying what it is you are hearing though.

[DorphonDalfir]

What do the minimum/maximum RMS levels denote? The dynamics between loudest/softest sounds? If so, are mine decent or too close together?

[Lee Knight]

I use Sound Forge and not Audition so I'm not sure. But with Sound Forge, those values would only be helpful for evaluating clipping issues or noisy bits of "silence". I'm not sure what value you can gain from analyzing those numbers in the context of a pop mix.

[alphajerk]

context is the biggest word in the post above. those numbers are meaningless without context. if you music is steady state all the way through, the numbers MIGHT mean something.... but if your average is -12db on say some dynamic and sparse music, its going to be a LOT louder tahn -12db on steady state music.

[Billster]

What the hell is "total" RMS anyway?

 

What I look at (and consider as a factor, but not as an absolute determination) is the range between peak and average dB. If the peak is going to be zero, something with a -18 RMS will sound softer than something with a -12 RMS.

 

The range I've found I like is something like -12 to -14 RMS with peak at 0, depending on the program material. I've analyzed some commercial releases (which sound like crapola) and found -8 RMS to a 0 peak, and there is no apparent dynamic - just a crushing onslaught of sound.

[chris carter]

I use Adobe Audition and I rarely rely on the RMS values. Case in point, I'll take records of mine that were mastered by the same big-time mastering guy and from the same exact album and I look at the RMS stats and they will be off from one to another by more than a dB yet sound the same.

 

Most mastering guys I konw use slow VU meters, which are only 'ballpark' anyway.

 

One trick I learned from someone was to use my trusty Radioshack SPL meter (analog style meter) and set it A weighted with slow metering. It's suprisintly accurate (provided your room acoustics aren't bonkers), more so than say VU meters working on the signal itself. The problem is that you can only compare relatives because of your monitoring level changing or where you hold the meter. But it does work VERY well to get me within a dB of where I want to be in a few quick seconds.

 

Ultimately though, there is nothing more accurate than your ears. RMS is only loosely associated with aparent loudness. Loudness is really determined by your ears.

[Lee Knight]

 

I use Adobe Audition and I rarely rely on the RMS values. Case in point, I'll take records of mine that were mastered by the same big-time mastering guy and from the same exact album and I look at the RMS stats and they will be off from one to another by more than a dB yet sound the same. Most mastering guys I konw use slow VU meters, which are only 'ballpark' anyway. One trick I learned from someone was to use my trusty Radioshack SPL meter (analog style meter) and set it A weighted with slow metering. It's suprisintly accurate (provided your room acoustics aren't bonkers), more so than say VU meters working on the signal itself. The problem is that you can only compare relatives because of your monitoring level changing or where you hold the meter. But it does work VERY well to get me within a dB of where I want to be in a few quick seconds. Ultimately though, there is nothing more accurate than your ears. RMS is only loosely associated with aparent loudness. Loudness is really determined by your ears.

 

 

Great post and nice idea with RS via the room acoustic. I'll try that.

[GZsound]

That is almost exactly what I do. I record both rock and roll and a lot of acoustic music and the percieved loudness between the two types of music makes RMS type measurements meaningless. I use Audition also..

 

I use my ears and VU meters on one of my passive monitor amps (an old Radio Shack reciever) to actually "see" where the volume really is after final mixing and mastering.

 

Using the level meters in Audition is nice for a guide, but can actually be decieving when listening to the "real" final volume.