Originally Posted by DonH50
The methodology is described in the papers (original and the update most often cited):
Fletcher, H. and Munson, W.A., "Loudness, Its Definition, Measurement, and Calculation", J. Acous. Soc. of Am., Vol. 5, 1933, pp 82-108.
Robinson, D.W. and Dadson, R.S., "A Re-Determination of the Equal-Loudness Relations for Pure Tones", British J. of App. Physics, Vol. 7, 1956, pp 166-181.
I'm aware of the papers but I haven't found anywhere to access them for free (my university logins for various journal aggregators have long been expired). I was lazily hoping someone might be able to provide a link to free copy or maybe summarise the methodology if they happen to know it (both pretty unlikely I know). The abstract doesn't mention anything beyond "new techniques have been introduced".
Originally Posted by DonH50
They do note that the curves are subjective, but are better than nothing IMO. I am not sure the variance among individuals is as large as you imply; what do you base your conclusions upon? That said, I know the proper relationship among SPL, sones, and phons is an on-going matter of debate since sones are indeed subjective. Most texts (at least the ones I have, lay and professional) do include disclaimers about the subjective nature of the curves.
In any event, this is not my primary field, so I am not really qualified to debate you or the writers of the papers.
I'm sure the papers are clearly outlining the limitations of the data as most respectable research papers do. But I suspect sound mixing staff (on movies etc) are implementing them with a rigidity and confidence that isn't warrented (what else are they supposed to do? I have no idea haha).
I base my conclusion on a small test I did with a friend a few years ago. Using a DSP and test tones, we each took turns to sit in the MLP of his system and perform the following steps:
1) turn off all auto EQ and other processing, using 'direct' mode on the AVR
2) Play test tone at 4KHz, adjust master volume till a calibrated SPL meter showed [x]dB.
3) Reduce test tone by 1/6 octave
4) adjust volume at that frequency using PEQ on the DSP* until it sounded
equally loud as the previous tone, ignoring the SPL meter (move your head enough to find a peak and a null to account for comb filtering)
5) Repeat step 3 and 4 until you get to 40Hz.
6) Make a note of all filters on the DSP and delete all filters
7) Repeat steps 2-6 with a different value of 'x' on the SPL meter.
*I can't remember what we used for Q-value on the filters.
We knew his sytem wasn't flat but it doesn't matter since we only wanted to see the differences between our results, it was quite large, as much as 6dB at some points, and usually 2dB or more (my sensitivity to 40-300Hz seemed much lower for example, but with a strong peak in sensitivity at 120Hz compared to my friend). I know its far from rigorous, I don't actually have the data anymore, and its only one data point, one or both of us could be edge cases for hearing but it does suggest...
I even repeated it by myself on my own system (far less capable) a few times and found that my own response seemed to vary across different days. I'd repeat it a few times in one sitting/day and get repeatable results, a week later; wildly different results. I realised it was mostly dependent on if I had been in a noisy environment (say at work) earlier in the day and also on my mood,
e.g.tension and anxiety would increase sensitivity above 1KHz. I guess I should try it again at some point, see if my hearing has changed in the intervening years (which it well might, I'm still a young man)
I guess I envisage a future auto-EQ technology that performs RoomEQ before some equivalent of this test for you a few times to create a more tailored frequency response for your individual equal loudness contour. Maybe even for multiple people just like current system use multiple measuring positions for RoomEQ. If anybody does this and makes money, I hereby claim 20% of all profit