Sounds good the bg noise is not bad at all, a nice clean recording overall.
http://www.affectbusiness.com/record/53.mp3I finally got these together. They're from the Squire Vallee-Vue Farm of Case Western Reserve University, east of Cleveland, Ohio, June 20, 2010.
I'll include only a few interesting tracks from the morning's recordings. This is not an "overly scientific" test for a number of reasons--I didn't take accurate distance measurements, and the two electret capsules are not identical. I'll guess that the parabola's is about 10 dB less sensitive than the one in the horn. The parabola is Edmund Scientific's 24 incher; the exponential horn is the one from sheet polyethylene that can be seen in a photo earlier in the thread, it's about a 11" square opening, 24" long.
Tracks 71 and 72 are in an open meadow with the recorder about 200 feet from a stream and an electronic musical instrument tuner about 100 feet from that. Looking at the sonograms, I can see that I was having a hard time holding the parabola accurately pointed at the sounds source. The focus is quite tight. The horn is looser, so the sound source can barely be heard at 90 degrees (the parabola's pickup at 90 is not evident on the sonogram, the horn's is). The horn gets much more of the stream noise when pointed away from the tuner than does the parabola.
60 (parabola) and 61 (horn) are a good pair from the same position, an eastern wood peewee at about 30 to 50 feet, about 30 degrees up. The red-eyed vireo singing is about 45 degrees off and is quite different with the two microphones. Wind noise in the background tree canopy is noticeably different between the two.
Track 53 is with the recorder's internal microphone, but with its software processing set on "zoom." Considering that this is a little handheld, it's really not bad (the recorder is an Olympus LS-10).
Track 56 is with an inexpensive hypercardioid microphone sold for camcorder use.
I'm not set up for formal testing--my last access to an anechoic chamber and related stuff was in Building 20 back in the day. The horn flares faster than it should, so by rights it should be oversensitive to bass. Still, it's directional, it's sensitive, birds spot it and fly as far as they can from it...seems like a well-functioning microphone!
I worked out what I think are the correct design equations given the error I made, and I'll try again. Doing it piecewise linear wasn't too hard, only cost my wife and me a few hours of arguing over the math. Figure out the differential equation for the curve was harder, and I knew I didn't have a prayer of integrating it. The Wolfram online integrator did it and was quite smug about it, reporting 0.03 seconds of computing time. Hmmph.
Maybe Saturday night I'll try recutting the gores, see if the shape really matters that much. I'll let everyone know.
I spotted this reference, I need to get to the engineering library to see if the issue is available (I can't find it online), or maybe someone here has old Institute of Radio Engineers journals?:
Clark, M.;
Bell Telephone Laboratories, Inc., Murray Hill, N.J.
This paper appears in: Audio, Transactions of the IRE Professional Group on
Issue Date: Jan 1954
Volume: 2 Issue:1
On page(s): 5 - 7
ISSN: 0000-0000
Date of Current Version: 29 January 2003
Abstract
An acoustic lens combined with a conical horn can be used to obtain a highly directional microphone without some of the disadvantages of the parabolic microphone. The directional characteristics can be calculated satisfactorily if one assumes that the horn provides uniform flooding of the lens aperture.
David
by Jake Hertz
by Duane Benson
by Duane Benson