Compensation curves that account for loudness

I've seen a lot of discussion about compensation curves, human hearing, and all that comes with it, but I haven't yet seen anyone talk about compensating for the fletcher-munson curve at the SPL of their measurements. I have seen discussions ln whether or not headphone frequency response changes with volume, and people have done measurements that show no differences as the volume goes up, concluding with "nope, no difference". But nobody mentioned the effects on loudness on human hearing.

So I want to do some experiments with this, as I have a plugin that applies the fletcher-munson curve to any audio that passes through it, making things sound psychoacoustically louder. You can vary how strong the effect is, and I've found practical uses for it too. It's very effective.

If anyone has REW .mdat files, or anything else that can be opened using REW, I can run those through the fletcher-munson curve and we can see what this does to the frequency response.

One ought to be able to make perfect target curves using this, where you can target a frequency response at a specific SPL. That idea excites me.

 

Or you know, just listen at @OJneg levels in the first place and not worry about equal-loudness curves .

 

Are you ancient enough to remember audio equipment with a button called loudness? This is exactly what it did.

 

Well, it kinda did the opposite. It applied a sort of reverse fletcher munson-esque curve.

I would love to see measurements with the appropriate equal loudness curve applied that corresponds with the actual SPL.

Edit: actually you're right, in that the type of curve it applied would be what I'd want applied to the frequency response!

 

Yes... the whole point was that you got to hear more of the music at lower volumes.

Basic U-shaped EQ is easy enough, but... you're looking for something more advanced, where the EQ changes with the sound level?

 

Well, if you look up equal loudness curves, you'll see what I mean. Human hearing scales in very particular ways as volume goes up. The "EQ" of your ears changes in odd ways. There are a couple dips that get deeper, and there's a peak that gets higher, and then the highs and lows roll off more and more in the more extreme ends of the spectrum. Frequency graphsthat are compensated for human hearing usually isn't compensated for this fact.

 

Yes, seen those curves over many years. That's not to say that I understand, but I do know of the concept.

I don't think they can be, because the SPL would have to be known.

 

Innerfidelity's graphs usually has the SPL listed, right? Regardless, you could just as easily do a rough simulation by adding the curve for the average "listening level"

 

All valid points. I just think it would be cool for the sake of curiosity. In some cases though, headphones will be a little hot in the 2-3k area, and you'll think "it's not that bad, probably not a big deal", but then if you look at an equal loudness curve you'll see that this area will only sound louder to you as volume increases.

 

Actually, they should have called it a Quietness button!

The phenomenon of the curves is not a getting-louder problem, it is a problem for low-level listening. You must have experienced this: there is point on the volume control below which you just can't hear all the music. It's not an audiophile thing, it is an everybody thing, which is why those loudness switches were ubiquitous on portable tape/CD machines. There are not for listening loud, they are for listening at low volumes. Of course, some people like that extra bass at other listening levels too.

I agree to bringing back tone controls. Ears are not equal; rooms are not equal, even all music sources may not be equal, let alone the differences in speakers and 'phones. Above all, tastes differ, and I no longer want to impose "high fidelity" on the world!

With increasing hearing loss, I wish I had not fallen for the amplifiers must not have tone controls line years ago. Mind you, that is largely irrelevant, now that the hifi/integrated-amp is barely used, and music is played on my computer where EQ is easy.

 

I always wanted a volume control with an internal DSP that applies an equal loudness curve that compensates for the volume level with EQ so it always sounds the same.

 

At Sonarworks we've played with such an algorithm. The problem is that your DSP needs to know the SPL's hitting your ears. Hence it needs to know headphone sensitivity/impedance and output stage capabilities.

 

The only way to implement it accurately would be to have a calibration routine that stores a profile for the individual amp/headphone combination.