[tuahogary]

I just did an interesting experiment in school recently involving the hearing of the human ears. It is known that our hearing peaks at conversation level or around 1 kHz to 8 kHz. We tested the threshold of our ears at different frequency and though I was not surprised at the result, it got me thinking about how one would perceive a "balanced" sound.
 
What happened was I had to wear a pair of headphones and listen to different frequencies and different volume to test my hearing threshold. For example, at moderate noise level (people around me chatting and all) my threshold for 7 kHz is a healthy 15 dB. But when the test got down to 750 Hz, my threshold is 45 dB. That is a massive 30 dB difference between what I can hear at different frequencies. I did not test it at higher frequencies but I imagine my threshold would be higher than 15 dB too. 
 
I know I may sound stupid when I ask this. My question is if a pair of headphones sound balanced to me, does that mean it is actually U-shaped in reality? I've seen frequency response graphs of headphones and are the flat ones really balanced?
 
Also, I have to admit it wasn't ideal conditions to test my hearing. For example, my brain's ambient noise masking may have affected certain frequencies but testing both ears individually and at regular frequency interval, I did get something like an inverted U-shape graph for my hearing threshold.

 

[jcx]

Loudness curves are well known in psychoacoustics - try search on Fletcher-Munson - basically if you really want "realism" you have to play back the music at the same SPL or the Loudness Curve effect will change the perceived frequency response
applying a compensating EQ (a "smile"  curve boosting lows and highs) can improve perceived frequency response balance when listening at lower than the live event SPL levels - but dynamics are not perfectly handled by a single EQ
 
headphone listening is different from stereo loudspeakers in a room in several ways so there is also debate on how to shape headphone frequency response to sound best with source material almost universally made for speakers
http://www.head-fi.org/forum/thread/550220/chesky-records-makes-a-high-rez-album-for-head-fi-ers-in-binaural leads to a good binaural demo
 
there is a general use of "free field EQ" for headphones as opposed to ruler flat in ear response to partially compensate for loudspeaker/room expected response - some headphones appear to have "built in" "Loudness Curve" boost as well as some that simply have exagerated bass

 

[tuahogary]

Hi thanks for the info! So is the Flecther-Munson curve supposed to be built into a headphone so that our brains can perceived the loudness as flat?
 
I also did not know that recordings are optimized for speakers. I have that binaural album you recommended. I guess you mean it is specifically recorded for headphone listening?

 

[inarc]

Neither microphones nor loudspeakers nor headphones need to compensate for equal-loudness contours, although headphones do have to simulate certain room acoustics. There is a difference between objective sound pressure and subjective loudness.

 

[jcx]

the usefulness of "Loudness Curve" EQ is to listen at SPL very different from the original with less perceived frequency response aberration
 
it isn't a inherent part of recording or playback but a nice feature to have on preamp/amp that you can adjust to your listening level to make listening more enjoyable when you don't want to blast your ears at live performance levels all day long
 
some headphones appear to have a "smile" EQ built in - like some Grados - making them "better" for listening to music genre that are naturally very loud at much lower average SPL from the headphone

 

[jimbob747]

Another very simple fact is that your headphones aren't acting as a direct source... So in ideal conditions the engineer creating the original 'mix' will be listening through 'flat' monitors and 'leveling' the sound to match. A flat, balanced or transparent sound is only true to the source which should have already been homed to human hearing (at least that of the sound engineer). I won't even go on about the differences in human hearing and how that'd effect how it sounds to him vs you.

This is however an ideal scenario, realistically in a lot of cases engineers will 'EQ' to suit the tastes of your common loudspeakers or standard 'stock' headphones, such as the ibuds so that the majority will enjoy the recording. This is one of the main reasons (alongside compression and all the rest) for the difference in quality between recent and older recordings as there was very little to change or 'master' using the old analogue format so detail and clarity always seemed paramount.

I wouldn't take my word as gospel as i'm just a school teacher who did a bit of sound engineering in college. I'm sure there are many (like those above) who have a much greater understanding and will give a better explanation. I just thought i'd throw in my 2 cents (if that's the phase i'm looking for).

 

[tuahogary]

Your explanation actually makes a lot of sense. Easy to understand and no audio jargons. Thanks. I've never actually thought of that tbh!