Am I reading Fletcher-Munson curves right

[Colin]

youth hearing at certain SPLs is very sensitive:

At 70dB: 330 to almost 1kHz, then again from about 3KHz to 6kHz

thanks

[lynnm]

I am not quite sure what you are saying/asking.

Loudness compensation (based on Fletcher and Munson's research ) is designed to boost the frequency extremes at low volumes to compensate for the fact that the average human ear requires those frequencies to be amplified in order to be percieved as being as loud as the midrange and presumably closer to correct.

At higher volumes the boost was felt to be unnecessary and even undesirable.

Is this what you are referring to or have I missed your point?

[Colin]

or how do you read the curves?

is there special sensitivity?

where?

[DrWho]

The curves on the typical graphs I've seen show you the SPL required of different frequencies in order to be percieved as the same volume. You could think of it as the opposite of the frequency response of our ears. The significance of the F-M curves is that they become flatter as SPL increases (though after 80dB it doesn't change as much). This is the main reason that louder music sounds better to our ears.

[mbajner]

I once had a Soundcraftsman 10 band equilizer that came with a record.

on the record were freq. associated with each band and the volume of each was based on the Fletcher-Munson curve.

was kind of cool to play with back in the 70's.

[Colin]

yes, but are we especially sensitive from 330 to almost 1kHz, then again from about 3KHz to 6kHz, at 70dB?

[DrWho]

Well at 70dB were just about +-3dB from the whole range 300Hz to 6kHz, but are most sensitive at 500Hz and 4kHz (which are midpoints of the two regions you specified). So yes, you're reading it correctly...

[Colin]

So, reading these curves again, I am struck my how much louder the low bass must be to be heard at normal volumes, and how much quieter the critical mid-range can be at almost any volume in order for people to hear particular frequencies.

At 70dB SPL, it looks like 100Hz requires about 5dB of volume, while 50Hz needs a large 15db boos and 20Hz needs an amazing 35dB! Three or 4 kHz on the other hand should be almost 10dB quieter than 1kHz. While frequencies from 7 to 10kHz need almost 10dB increase, the ones above that dont. We seem to sensitive to some frequencies above 10kHz. In fact, looking at the curves, I am not sure where this notion of 20-20kHz came from. Our hearing is not flat at all. Perhaps loudspeakers or tone controls should be made to fit these curves: dial a SPL level and that is the attenuation you get at each frequency to output an ideal response not flat, but matched to our hearing ability.

The curves never really get flat. We always hear the same frequencies better, even when they are louder. But according to the curves, I should not hear much below 300Hz at 70dB and I know that is not true.

[michael hurd]

That is the problem with designing a speaker that measures flat. ( or relatively flat ) At low volumes, it seems as if the bass is lacking in respect to the midrange and high frequencies.

At higher volumes, you notice that the speaker does indeed have sufficient bass. At what volume do you get to listen to speakers in a retail setting? IMO, a speaker with a big knee or hump in the bass, in a showroom will get more attention than a more accurate reproducer beside it.

With the advent of tone controls on preamps and integrated amplifiers, and more powerful equipment, speakers could become damaged through the misuse of the bass and treble adjustments. Thus the manufactuers slowly moved away from adding these controls.

On today's modern audio/video receivers, we see that the tone controls are again in vogue. Most of my serious listening is done at 85 - 90ish db, if it is background music low in volume, I will add 2 - 3 db of gain in the bass to kind of make up for the perceived lack of bass. Of course for most listening, I select direct mode, which bypasses the tone controls altogether.

[colterphoto1]

Interesting topic and DrWho has once again hit it on the head.

I think a philosophical discussion of what the human ear was designed for should accompany this. Not to mix Creationism with Darwinism, but either might apply here. Either we were 'designed' to hear the human voice and sounds of nature, or evolved that way, perhaps a bit of combination. If you couldn't hear the wolves walking through the forest floor or the trample of elephants you were eaten. The 'survival of the fittest' theory is sometimes related to the ability to procreate, so you'd better be able to hear the female of the species' 'sweet nothings' in your ear if you were to uh, get any. LOL

In any case, the 300-6K, which I think is not just what is more sensitive in early life, but whats left after 50 years of abuse is pretty much the human vocal range and sums up most natural sounds. Perhaps we weren't designed to hear subs sliding into the abyss, hurricanes, or the high shrill wail of synthesizer leads. So crank up the eq's gang!

Just a bizarre thought after a sleepless night.

Michael

[Arkytype]

Colin,

Print the F-M family of curves then invert the page top to bottom. Looking thru the page, 20 Hz should be on the upper left frequency scale and the lowest SPL curves will be at the top. Now you can get a better picture of the human ear's "frequency response".

As you can see, human hearing is not "flat" at any SPL.

A question for all you budding recording engineer out there: at what SPL level should you mix a multitrack recording to a two track mix and at what level should the consumer listen to the recording in his or her home?

Try this site to test your hearing. Haven't tried it, but it looks interesting.

http://www.phys.unsw.edu.au/~jw/hearing.html

Lee

[ZAKO]

Believe me....Recording engineers are very familier with the F.M. curve.

[Colin]

cool site, but yes, that is what I did, rotated the curves, still haven't played with the DEQ2496 much yet...

[DrWho]

----------------

On 2/10/2005 12:34:12 PM Arkytype wrote:

A question for all you budding recording engineer out there: at what SPL level should you mix a multitrack recording to a two track mix and at what level should the consumer listen to the recording in his or her home?

----------------

Every textbook says between 80 and 90dB, but I prefer to reference all my mixes before the mastering process. I'll listen in the car, at home, on the pc, at friend's houses and so on and so forth. I then make decisions based on trends found between all the systems at the volumes I found to be natural at those times. Then in the studio, turning the music down is a great way to detect certain problems whereas cranking it up to like 110dB brings out other things. Nobody listens at the same volume anyway so I make sure it sounds good (or as good as possible) at all of them. I must confess though that this takes bloody long to accomplish, so I only worry about it when producing more important stuff. I don't have enough time to my disposal.

[Arkytype]

90 dB sounds about right. Unfortunately, those of us who mixed rock and roll in the wee hours of the morning in the '70s may have started out at 90 dB. After an hour or so, the amps were clipping and nobody noticed. Must have been the medicinal herbs that were prevalent back then!!

Recording engineers wanting "real-world" speakers would buy the single 4-inch driver Auratones and set them on the edge of the meter bridge. If you can hear all the instruments on these, then your mix was done.

Lee