SPL-flat equalization target for between 20 and 1000 hz
Jul 18, 2019 at 8:10 PM Thread Starter Post #1 of 3

AutumnCrown

100+ Head-Fier
Joined
Sep 24, 2015
Posts
262
Likes
71
I need feedback on a possible method of estimating loudness of headphones. My goal is to make estimating volume levels doable for anyone with access to headphones which can be equalized to a given target.

The methodology is basically to create two tones with the following generator, one at 30 hz, and one at 1000 khz, with the same SPL level.

https://www.audiocheck.net/audiofrequencysignalgenerator_sinetone.php

I then adjust volume until I can barely hear the 30 hz tone. Then, according to the contour, and assuming my hearing is undamaged and relatively normal, the 1000 hz tone should also be about 60 decibels, and repeating the test with a 20 hz tone should leave the 1000 hz tone be at about 70 decibels.

My problem is I am not sure what target response of the headphones should be equalized to. I assume it should be somewhere around the Harman target, but then again, perhaps the diffuse field bass response would be more "spl-flat" between 20 hz and 1000 hz?
 
Jul 19, 2019 at 5:07 AM Post #2 of 3
Why don't you try it and see what happens? However, I would very strongly advise that you do it the other way around: Set a level at 1000Hz, then switch to 30Hz and increase the output volume. Be EXTREMELY CAREFUL to lower the volume by a huge amount if/when switching back to the 1000Hz tone from the 20Hz/30Hz tone!!!

G
 
Jul 19, 2019 at 12:22 PM Post #3 of 3
I need feedback on a possible method of estimating loudness of headphones. My goal is to make estimating volume levels doable for anyone with access to headphones which can be equalized to a given target.

The methodology is basically to create two tones with the following generator, one at 30 hz, and one at 1000 khz, with the same SPL level.

https://www.audiocheck.net/audiofrequencysignalgenerator_sinetone.php

I then adjust volume until I can barely hear the 30 hz tone. Then, according to the contour, and assuming my hearing is undamaged and relatively normal, the 1000 hz tone should also be about 60 decibels, and repeating the test with a 20 hz tone should leave the 1000 hz tone be at about 70 decibels.

My problem is I am not sure what target response of the headphones should be equalized to. I assume it should be somewhere around the Harman target, but then again, perhaps the diffuse field bass response would be more "spl-flat" between 20 hz and 1000 hz?
my feedback is going to be that it's not a good idea. I get the idea and if a bunch of assumptions are correct then indeed you'd be in a close enough ballpark of known SPL. except that each assumption is only that. the equal loudness contour is an average so already that's a problem. the measurements were done in an anechoic chamber, so unless you're confident about the ambient noise level being low, it might in some cases mask the just noticeable amplitudes, or at least impact your overall impression of loudness and still make it harder to notice the quietest signals.
then a headphone has to measure flat down to 30 or 20Hz, otherwise it doesn't mean much. but even more important, it has to have low distortions at those frequencies at such output, otherwise you might mistake some THD at higher freq when the 20 or 30Hz itself is already not noticed.
all in all, you'll be likely to accumulate small errors here and there with a final result that's going to be pretty uncertain.

anybody looking to estimate loudness would probably get a more accurate result by trying to know the output voltage of any device they own at a certain output(maybe ask online if someone can measure it into a given load), and then rely on the sensitivity spec at 1kHz provided by the manufacturer of any IEM or headphone you have. that might still result in errors, but it will usually come quite close(some manufacturers are known to often give bogus specs, but it's relatively rare to have big differences).


of course I'm looking for what could go wrong instead of being optimistic, your idea could lend you close to some correct SPL value pretty often with a little training. but I've discovered that life has a way to make the "what could go wrong" happen.
 
Last edited:

Users who are viewing this thread

Back
Top