[LAD] twice as loud

Paul Davis paul at linuxaudiosystems.com
Thu Jul 22 20:57:41 UTC 2010


On Thu, Jul 22, 2010 at 4:50 PM, Philipp Überbacher
<hollunder at lavabit.com> wrote:

> We may be comparing the wrong thing when we compare with the size of
> objects to loudness.
> It's relatively easy to say that the interval between sound B and C
> is twice as long as the interval between A and B (given the
> interval and the length of the sound is in a certain range). This is
> probably closer to the object size comparison.
> I wonder how well we can judge something like twice the
> brightness.

one little side problem with this is that our sensitivity to both
loudness and brightness is adaptive. this means that although one
could do some experimental work to determine the ratios that lead most
people to judge one sound 2x as loud as another, as soon as you leave
the experimental context, it becomes pretty meaningless in any
practical sense. what you judge as quiet or loud (or bright or dim)
depends an awful lot on what you've just been listening to. given that
our sensitivity to volume is non-linear, it only takes some
pre-exposure to a very quiet or very loud environment to totally skew
the part of the curve that we're on when we try to establish how loud
something is.

to be clear, i'm not suggesting that its not possible to come up with
some useful and interesting numbers by measuring this sort of thing. i
just want to note that they have to be viewed as deeply fuzzy because
of the effect of the pre-listening environment in setting sensitivity
levels.

--p



More information about the Linux-audio-dev mailing list