Fletcher-Munson curves: Applied uses and lessons learned?

Question

There are a few posts about the Fletcher Munson perceivable loudness curves on here, more explaining what they are but not their function. I haven't quite found out if they have a function or are directly relevant or applied to anything of use.

My thoughts are that they might be used as a tool to help, for lack of a better word "hack" your way to a better mix, as that's ultimately the end goal.

Could the curves help inform or hint at ways you could boost the perceivable loudness of a mix, as if you were mixing it to fit into the ear best?

The parallel I can see just by visually studying the FMC, is you are most sensitive in the 1-4k region, and that's already something I do when mixing, try to take out the harsher sounds so you can push turn it up louder without fatiguing.

Are there any other bands that are important to notch out or any other lessons that can be learned from these curves?

Answer 1

Fletcher-Munson curves, or more generally equal-loudness contours, are mainly used when measuring sound pressure levels and noise. The definition of such curves roughly goes like this:

An equal-loudness contour is a measure of sound pressure (dB SPL), over the frequency spectrum, for which a listener perceives a constant loudness when presented with pure steady tones. (source: wikipedia)

So with that in mind, I'd say the tendencies of the curve can help you produce a more even mix with regard to frequency distribution. And the curve may help you focus your efforts on dynamics in frequency areas we are more sensitive to.

Equalization and frequency distribution

Let us pretend you have a mix with a variety of sources in different frequency areas and/or sources that individually cover a wide frequency range. The task could now be to level and equalize these source so the outcome over time approximate the fletcher-munson curve (around 40 phon).

The way you monitor this is with a real time spectrum analyser that is set to maximum time window (as slow as possible, to get an average of as large a regions as possible, and thereby minimize temporal deviations), or even better, perform a complete analysis on the exported program material (for example using Har-Bal or similar tools).

Dynamics

I see two different angles on the dynamics application:

1. If you look at how the curves vary at different sound pressure levels, you may be able to use these observations to control a dynamic EQ, so that it compensates for the general RMS variations throughout the song.
2. If you look at one of the curves (pick the 40 phon curve) and note the variations, you can use this to pick regions for a multiband setup, and compress harder/lesser in regions we perceive more quietly (I'm not certain what really would improve things here.. help me out)

---

In general though, when mixing I think you are better of focusing on the individual sources and balancing the mix without analytical regard to these curves. There is a hidden premise in all this - things will sound better if you approximate the curves - and I simply don't think that is true. Various instruments and voices just sound differently. You will never get that piccolo flute to align the lower regions, the overtone profile may get completely distorted if you deliberately try to force out areas that the instrument dont cover etc.

In a mastering context it may make sense to approximate the general profile towards the 40 phon curve, but I would never do this blindly as there always will be frequency ranges in the program material that simply are not as active as others due to the nature of the sources. Hence the advice on using a large window / average region (perhaps even several songs / a complete production).

Answer 2

My thoughts are that they might be used as a tool to help, for lack of a better word "hack" your way to a better mix, as that's ultimately the end goal.

To clarify some things : There's no hack in mixing.There are techniques that get you there faster, but hack - no!

Also you can't just notch 2 frequencies and expect the perceived loudness to be insane. Of course it does make a difference making a V shape on your song EQ taking out some mids and boosting hi's and lo's. Also there's a classic mastering technique the crossover mastering(there could be other names for it too) which helps with loudness and control.

Keep in mind that no matter what frequency the human ear takes at loud volumes , the first frequency region that you start to "lose" or the ear itself starts to smear , or others say compress , is 4K which of course comes back after a while , but it's the first line of defense the ear has, to drop 4K. Also middle aged people that start having some hearing loss , also lose 4K first.

Sibilance and presence lies before and after 4K which are very usefull when trying to make something loud.

The parallel I can see just by visually studying the FMC, is you are most sensitive in the 1-4k region, and that's already something I do when mixing, try to take out the harsher sounds so you can push turn it up louder without fatiguing.

What you state here , is that actually you want to FIGHT the purpose of loudness, you want to be able to turn up more , without it being so loud.

Loudness's purpose on the other hand , is to be as loud as possible.

So if we have a guy with a nice amp at his house , he wants to listen to some loud music , he plays your song and turns up all the way up to 7. 7 for him at that point is loud music he likes it!

Now he goes and plays another song with all the tricks that make it be perceived as louder , and turns the volume down to 5. The music is still loud for him , he doesn't even care why he turned it down.

Loudness is not a tool to let you push more when listening, it's the exact oposite.

It's an actual dogfight, a way to stick out when the radio plays at the same volume , or to drag the listeners ear when playing a list from spotify , ears like LOUD!

"hack" your way to a better mix

A louder mix, is NOT a better mix.

Loudness is only a fraction of what does something sound right , but sometimes engineers start their whole mix , in order to end up boosting the loudness , which is mixing for a purpose. You can't just go with a loudness curve and fight with those guys, you have to start building everything from the start , so the loudness curve at the end is more "efficient".

Mixing is a series of serious decisions, deep applied technical knowledge and a world of obstacles to fight, from bad rooms and speakers , to apple's mono output cause hey , guys now listen to music from the iphone speaker !?!?!?!

If you want to be loud , start accepting the idea that you'll lose a lot , musicality , harmonical content , dynamic range and others that make a song have soul.

Good luck.