Fletcher-Munson Curves

Is our perception of bass really this bad? Fletcher–Munson curves - Wikipedia

 

Yep. That's why the loudness button on the reciever works by just amplifying the Bass and treble without increasing the volume of the entire signal. That makes everything sound a lot louder. It's also why mastering engineers insist on working at exactly 85 dB, because the sound signature will change at different volumes.

 

So, adjusting the volume so that 85dB is registered on a sound level meter at the listener's seat should give the optimal volume level for listening!

 

Two things. A loudness button or switch will only result in "accurate" compensation for a given amplifier output, speaker sensitivity and listening distance. Yamaha had a variable loudness control, which seems like a better solution, but one you'd have to adjust by ear. Also, I thought I'd read years ago that the Fletcher-Munson curve was erroneous in that while it is true that the sensitivity of human hearing is reduced in bass frequencies, it is not so in the high frequencies. Perhaps mistaken lab results based on equipment limitations at the time (1930s).

 

In theory, if you want to hear what the mastering engineer hears. It's also the recommended level for movie theaters and is considered an optimal "moderate" volume. Some people adjust it to about the volume to what a specific instrument in the mix would normally project.

 

Instead of a loudness feature, when listening quieter like I do, you can implement your own loudness curve, also known as a house curve when integrating a sub in your system. I listen about 75-80db. 85-90db is for rocking for a song or two in my world but quickly gets old, even with alcohol. I'm old.

So I built in a house curve when tuning my sub. Started with flat to 100hz and it doesn't work for me at my listening level. My curve at 40hz is a few db hotter than the mains at 100hz, when measured at 80db.

 

I 've known about the F-M curves for some time now, but until today, I never realized the actual amount of bass 'roll-off', and never noticed how big it was compared to treble.

 

Equal-loudness contour - Wikipedia

Reading the above link you can see that this is not an exact science. Couple that with the fact that we all hear differently to varying degrees and I would suggest that whilst they may offer some rough principles you shouldn't get carried away with them.

In other words, make adjustments that suit you.

 

The Fletcher-Munson curve accounts for why Dr. Dre's Beats headphones are such a brilliant idea, and such a popular consumer product. The bass boost of those headphones addresses the psychoacoustics issue. Granted, not in a way that allows for fine adjustments, but in a way that provides a substantial improvement in listener satisfaction over traditional flat-response headphones. The same way I've been known enjoy hitting the loudness button, when listening to music at low volumes.

 

The point at around 3kHz, where we perceive sound to be the loudest at any volume level, is the same frequency exploited by a baby's cry and alarm systems. That's why they're so grating and so galvanizing - it's an evolutionary development and a means of getting us to move, respectively, to survive. Not necessarily relating to the topic at hand, but interesting nonetheless.

 

That theory assumes that listeners will listen at safe levels. Or do Beats reduce the curve as volume increases? My guess is not.

 

Interestingly, it's also around this frequency that hearing loss is usually greatest due to loud noise exposure.

 

I do not use this with truly full range setups.
The reason is that at abnormally low volume it is just as valid an idea to "shrink" the sound.
And on a full scale set it is nice to turn it down and "see" the image shrink.
Adding a bunch of boom to it just adds mud.
The last time I needed a loudness curve was with an underpowered boring mid-fi system that lacked drama.
Loudness compensation, like adding compression, is mostly an effect useful for a car stereo IMHO.
When listening ten hours a day at home sometimes shrinkage is nice when you are not in the mood for wall to wall sound...
A full coverage system should be able to do more than one trick.
But as with everything else if it makes you happy happy---do it!

 

I blame babies and sirens.

 

and also why many speakers have a dip in their response in the 3K region.

 

OK. So if people working in the studio and doing the mixing and mastering all work at 85dB with really flat headphones and monitors, they've tamed a relatively mild F-M curve. Loudness buttons can help with playback at lower volume. But why are speakers/headphones with u-shaped responses so popular? Are they mostly appealing to folks who listen to music at lower levels (<85dB) or is heavy bass and treble sizzle just a current fad?

 

No, the Fletcher-Munson paper was like 85 years ago, and audio research has moved on a bit since then. The updated/more accurate ones are now referred to as the Equal Loudness Curves.

 

I hate those loudness controls. Just train your ears to compensate for it

Do you have a source for your 85 dB figure? Not doubting it, I just want to read more about it.

 

Plus, isn't the reference level just that- something for calibration using test tones? I would think listening at a constant 85 db (not accounting for peaks) would be very loud, thus you often see -db (from reference) as a measure.
In practical terms, my system, which is fairly substantial in a pretty large room, delivers full range at less than 80db at the listening position (roughly 13 feet from the front line of the speakers), and thus even further below that measurement at say 1 meter from the speakers.
There's no question that at low volumes you aren't getting the full range, particularly in the bass. But even with refinements to the loudness curves over time, I often wonder how much these measurements take into account noise-- which masks sound. Ambient noise in the room and environment is a factor; another is system noise--because I use extremely efficient horns, I was forced to reduce the noise floor of the equipment and system overall through component choices, the manner in which wall power is supplied to the system, and inter component grounding, all of which you could hear if you turned up the volume on the system without music playing. Once I achieved a level of quietude, your could hear more information at a given loudness. So, for me, having a highly efficient speaker system isn't just about more db/watt or the ability to play the system at crazy volume levels, but the associated need to reduce the noise floor, which yielded a quieter background against which the music is presented.
I'm also pretty sensitive to loud noise and ambient noise, e.g., as I have aged, it is more difficult to hear conversation in a crowded noisy restaurant.
77 db averages at "C" weighting on a decent SPL is plenty loud for me- on occasion, for heavy rock, I will goose it up a little, but taking into account peaks on dynamic material, for me, my ears, my system, my room, this is plenty loud.
I think a lot of people would be surprised at the SPL levels if they measured using a decent meter. We were in a "trendy" restaurant not long ago here, and it was insanely loud- I happened to have a db meter app on my phone- perhaps not the most accurate, but I was getting +80 readings at our table. By contrast, my listening room is around 37 db without music playing, which is pretty quiet in a house in the city.

 

The reference level depends upon the room size too, 85dB SPL in a small untreated room is gonna sound horrible and way too loud. 85 in a large treated mastering room, film sound stage, or cinema, is gonna sound pretty great, and indeed, if your cinema is a good one, that's the level it will calibrated to.

For mastering in my medium sized room I go with 77dB SPL as the reference level.

 

I originally learned about listening levels in a college class I've been taking. The book we're using is "Modern Recording Techniques 8th Edition" by David Miles Huber. Chapter 2 reads "If a piece is mixed to sound great at a level of 85-95 dB, It's bass and treble balance will actually be boosted when turned up (often a good thing.)" Another quote reads "Over the years, it has generally been found that changes in apparent frequency balance are less apparent when monitoring at levels of 85 dB SPL."

 

You just have to look at the curves. The "flattest" curve in the frequency domain is when we listen at around 85dB SPL.

 

Looks to me like the 100 dB curve is flatter in the bass and midrange, especially flat in the octave centered around middle C, but we're not doing all our listening at 100 dB average volume, at least not at home.

The problem for the listener is, unless every performance level is chosen by the performers with a specific known test volume in mind, say, 77 dB at 1kHz, or whatever. And then every recording is monitored and mixed at that level. We don't know what is the "right" reference level to set our systems at relative to what the performers and engineers did, and of course without that standard it's going to be different from performance to performance and recording to recording.

Plus, we all know, psychologically, what music sounds like when it's quieter or louder, or when we're closer to or farther away from the instrument and instrumentalist. So, we kind of reflexively make a psychological adjustment when we hear music. And so to do musicians and instrument makers. We've never heard music other than with our non-linear hearing. We know what it sounds like at different volumes. And music is played for and instruments are designed to be heard by humans playing sometimes loudly sometimes quietly.

Furthermore, I think we've all had the experience that our systems in our room have a volume level at which they kind of snap to life, at which the instrumental scale seems to start to become more realistic, at which the acoustics of the recorded space seem to begin to best reveal themselves, etc. Is that a matter of listening room acoustics, levels at which the electronics are best driving the speakers, levels at which our listening level is more closely matching the frequency balance the engineers heard because of these non-linearities in our hearing? Probably all of those are playing a role in that experience.

That's why I don't really sweat the equal loudness bag -- too many variables. I just know music sounds one way at one volume, and another way at another volume. But I do like listening to my hifi best when I can turn it up to that snap-to-life volume.

 

I don't know that speakers have a 3kHz dip in response to non-linearities in human hearing.... I'd love to hear from any designers if they're actually working around taht. I think speakers tend to have a dip off axis in the region where a woofer crosses over to a tweeter, which, with a lot of these common 1-inch tweeters, tends to take place in the 2k-4kHz octave.

 

And is seems many engineers BOOST the loudness level around the 3kHz ot 4kHz level as to make me NOT want to listen to their music. I'd rather listen to a baby's cry or a car alarm.