Let me try and describe what is going on. It may take a separate post
and starting over from scratch as I agree with a couple off list
e-mails that I received. Those indicated my photo and description
both sucked and was useless unless you already are doing a bunch of
parabola stuff along these lines.
This is a parabolic dish with the mics separated enough to give
separate and unique focus points to each mic capsule. Sounds that
appear the same volume to both mics would not be in focus to either
mic from the gain of the parabolic dish. Therefore, if you invert one
channel and sum the two channels to being mono those off axis sounds
get cancelled out. What is left is only the sound heard as reflected
from the dish with gain at either of the mic capsules.
The more I think about the math of it and actual gain or noise
reduction the more trouble I get into. An example is easier to show
that the concept works than trying to explain why or by how much.
In the example sound file, those popcorn like sounds are small millet
seeds being broken by a small bird at 100 feet and the wing flaps are
mainly little 5" birds as well. The recorder that captured these
sounds was a consumer sony md recorder and the dish is a 32".
I am recording tomorrow and will try and get better examples and
photos. Because Canada loaned Minnesota all of their owls for a
couple months I have to give it a shot to record Northern Hawk Owls
while they are here in MN in good numbers. After this winter I
suspect that homeland security will have made the border so tight even
a hawk owl couldn't vacation here without papers and an interview. :)
Rich
--- In Rob Danielson <> wrote:
> Neat Trick. I think the 183's are sounding pretty nice. So is the
> theory the dish shapes frequency response and these frequencies are
> getting phase cancelled out or is it an ampltiude difference? Can you
> run through how this works again? It seems like part 3 shows
> something like a -6dB roll-off starting at about 700hz. The distant
> machine is centered around 240 hz. Do you think pt 3 is effectively
> cancelling more in this Hz area? For a stereo result, attentuation in
> the 240Hz area seems to help bring out the spatial image and improve
> overall clarity. Your caps show a similar zizzz noise concentration
> ~8KHz-- maybe a few hundred cycles higher than mine. Rob D.
>
________________________________________________________________________
________________________________________________________________________
|