Why is it that a certain song or wave file will have a set limit where it's Average RMS Level cannot become any louder via Limiting or Compression?

For example, playing around for fun, I would use a 60 second section from a song, and use a Limiter, usually the Waves L2, I would keep lowering the Threshold until the Average RMS Level would actually start to go down(become quieter) again.

For example:
THRESHOLD = Average RMS

-10.9 = -4.9 dB
-11.0 = -4.8 dB
-11.1 = -4.8 dB
-11.2 = -4.8 dB
-11.3 = -4.9 dB
-11.4 = -4.9 dB
-11.5 = -5.0 dB

So for this example, this particular song can never have an RMS Average louder than -4.8 dB. But the question is how or why?

I don't understand this at all? I'm just trying to learn as much as possible about Compressors and Limiters and how they work etc...

THANKS:)

RMS can be calculated in a variety of way. Two variations are relative to a sine wave and relative to a square wave. The differences here are about 3dB. If you compute relative to a sine wave, the RMS result is about 3dB lower. That might be part of what you find.

If you are working with integer values, e.g. 16 bit, there is an absolute maximum sample value that can be encoded. This will always be the peak, not the RMS, although perhaps the square wave approach comes close; I'm not too aware of all the math involved. However, it is also the case that, depending upon the particular wave form, the peak output value can be about 8dB above the highest sample value (although such extremes are, at best, very rare in music). This is easy to see in a program such as Audition that displays the wave form as a good representation of what a proper DAC will produce. I suspect the RMS calculation does not take that into account.

I think you might be thinking too hard on this...is it really that complicated? Maybe I'm wrong...

But I think it's obvious that no one can take a song and somehow use a Compressor/Limiter to such an extent that it will eventually reach an average RMS volume of -0.0 dB. Obvious yes, but how to explain why it's not possible is what I'm really looking for.

In other words, you can't just keep increasing the Limiters Threshold from -10.0 to -10.1 to -10.2 to -10.3 etc..etc..etc...because eventually the RMS volume of the song will no longer continue to become louder, it will eventually start to go DOWN in volume again...always...but why? I don't know!? That's what I'm trying to figure out...

RMS isn't psychoacoustic in any way, so while it can become louder it's certainly possible to reach a limit in artificial measurement... but maybe it's also the way you're running the limiter, a limitation of it. I'm not sure exactly how RMS works, but I'm sure it wouldn't take a very accurate picture of the volume of a completely-limited sound - where every sample is at 0db, alternating positive and negative.

Reminds me of when I used to louden .mod samples by doubling the volume repeatedly.