On 01/22/2009 05:47 AM, Maik Justus wrote:
>>> Just to clarify on the reference-dist, is it that this value is a
>>> diminishing effect, that is for reference-dist of 1 after distance 1
>>> the volume is half original, after distance 2 the volume is 1/4
>>> original (half of a half), distance 3 it's an 1/8th (half of a
>>> quarter).
>> yes, exactly.
>>
> not exactly, it's 1/8th at distance 4 (doubled distance result in half
> volume).
Sorry, no, it's not any of those. In the present situation, the
levels go like this:
r power power
/ r0 / (w/m^2) / dB
0.5 4.00000 +6.0
0.7 2.04082 +3.1
1 1.00000 0.0 <<< reference
1.5 0.44444 -3.5
2 0.25000 -6.0
3 0.11111 -9.5
5 0.04000 -14.0
7 0.02041 -16.9
10 0.01000 -20.0
15 0.00444 -23.5
20 0.00250 -26.0
30 0.00111 -29.5
50 0.00040 -34.0
70 0.00020 -36.9
100 0.00010 -40.0
We see that at the reference distance (r0), the signal is not
attenuated at all. That's the defining property of the reference
point. At twice that distance, the signal is down by a factor of 4.
At three times the distance, the signal is down by a factor of 9.
It is the famous "one over r squared" law. It is a corollary of
conservation of energy.
*) At larger distances sound energy is not (by itself) conserved, i.e.
dissipation becomes dominant, and we see a crossover to exponential
attenuation, but ...
*) At the distances we see in flyby view, dissipation is negligible.
The 1/r^2 attenuation is the whole story.
If you know the sound level at any one distance, you can calculate
it at any other distance.
On 01/21/2009 05:46 AM, James Sleeman wrote:
>> ... if we switch to tower view, it seems you can always hear the
>> aircraft no matter how far away you get, for example, I was 100 miles
>> from the tower and yet I had no trouble hearing the aircraft at all.
That's a bug. The tower cab has lots of sound insulation, so the
tower guys are not going to hear the aircraft at all unless it is
very close. If it's not close, 1/r^2 attenuation predicts that
the sound level will be inaudibly low. And dissipation makes it
even lower.
On 01/21/2009 05:14 PM, James Sleeman wrote:
> It seems that a great many aircraft do not define
> these values at all. Is there a default definition for these somewhere,
> is one calculated by openal maybe in the absence of these specific settings?
IMHO it would be a step in the wrong direction to ask aircraft designers
to specify the reference distance. There's already a length-scale
built into the flyby view, namely the expected distance of closest
approach.
There needs to be some headroom in the sound level, because the aircraft
might maneuver so as to come closer than expected.
> On 01/22/2009 03:17 AM, Vivian Meazza wrote:
>> I would think that the attenuation of sound in air is amenable to
>> mathematical calculation.
It is. In the near field it goes like 1/r^2. In the far field it
is exponential; see FAR A36.7 if you want the lurid details, or
see http://www.sfu.ca/sonic-studio/handbook/Sound_Propagation.html
if you want something more explanatory.
>> Surely we shouldn't be guessing at some arbitrary
>> "reference distance"?
There should be no guessing involved ... but there does need to be
a reference of some kind. There needs to be something to set the
scale. This is the premise of the statement above:
_if you know the sound level at any one distance_
you can calculate it at any other point.
On 01/22/2009 06:05 AM, Melchior FRANZ wrote:
> But it depends on the frequency pattern, no? So we'd need to
> analyze the spectrum ... time to use libfftw3.
No, the 1/r^2 attenuation is independent of frequency. No FFT
required.
The long-range exponential dissipation would be another story, but
we don't need to go there, not for the applications presently
contemplated.
------------------------------------------------------------------------------
This SF.net email is sponsored by:
SourcForge Community
SourceForge wants to tell your story.
http://p.sf.net/sfu/sf-spreadtheword
_______________________________________________
Flightgear-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/flightgear-devel
