50 years later, how would you improve this.

It is pretty easy (and cheap) to get an OpAmp that will replace the Lafayette amplifier.

One could go "hog-wild" and put an electret microphone on each tube. Phil Hobbs would be the proper person to tell us if there would be a noise (or spectral) advantage and how to sum them altogether... I wouldn't have a clue. ...Jim Thompson

Note that the multi-tube contraption is not really called a shotgun microphone. This is what the industry now calls a shotgun mic: Howevery article I can find also calls the tube variety a shotgun mic, so I guess both types now share the name:

I built 2 of those in college. The theater arts department continued to use them for about 10 years. There's plenty of room for improvement.

In the real shotgun mic, there is some foam between the tubes to mechanically isolate each tube. That prevents coupling between tubes to help flatten the frequency response and help to reduce "whistle" effects.

If you put one (electret) microphone in each tube, you could:

1. Compensate for group delay (phase shift variations between tubes) in frequencies that overlap in adjacent tubes.
2. Insert a band pass filter after each mic to eliminate sensitivity to harmonics.
3. Compensate for variations in gain between the tubes and mics to obtain a more level frequency response.
4. Perform common mode cancellation of sub-harmonics to reduce side lobes.
5. Perform common mode cancellation on anything that is mechanically coupled to the tubes.

Or, perhaps start over with an acoustic lens?

The multi-tube shotgun mikes were popular until people realized that you get the exact same effect with a single tube that has a slot down the side, where the width of the slot is a carefully controlled proportion of the bore size of the tube. The slot acts like the "open ends" of many tubes of graduated lengths, but you only need one mike.

I believe that most professional directional mikes are now built using a single tube. You can probably find details in the patent literature.

Clifford Heath.

I haven't studied the physics. But the idea came from folk who were drilling a number of individual holes along the length, to create resonant peaks, like a flute or recorder... but you want a flat response so you have an infinite number of holes an infinitesimal distance apart... IOW, a slot.

The slot width is obviously designed to match acoustic impedance (energy traveling down the tube vs energy exchange with the outside air), but I can't do the math.

The details are guarded by the manufacturers, of course.

Clifford Heath.

Oughtn't the free-air sound velocity and in-tube sound velocity be slightly different? If so, one can tune the group delay by rearranging the tube bundle to not have all output-ends coplanar.

Another bandpass possibility is to put a perforated stop over the tube entries; this makes each tube a Helmholtz resonator, and lowers the resonance (you can shorten the tube) while raising the Q.

Helmoltz resonator or bandpass filter, any phase shift will interact with the group delay correction.

1. try materials like PVC pipe or titanium (boiler tubes are semi-affordable)

Aha!

Is the slot constant-width? ...Jim Thompson

Neat contraction for "however every"

article I can find also calls the tube variety a shotgun mic,

Yes, slim pickins trying to find the type I was interested in.

Great you found two I didn't.

When you said could: did you mean, probably would need to:

Handy:

Could:

I need to read that again, it didn't compute the first time through.

Mikek

I don't think an optimal implementation would require 37 tubes. One strategy might be to match the tubes to the critical bands (

Especially if combined with a single microphone per tube, this might outperform the slotted-tube approach that Clifford mentions.

-- john, KE5FX

It's much easier to do with an ambisonic microphone - google it. You only need four accurately matched sensors, and with a bit of math you can calculate the direction of every sound source.

Dunno. I would guess(tm) that Bernoulli fluid effects would slow things down near the tube, but I have no idea how much or if it's significant.

This might help: "How A Shotgun Mic Works" "The more tubes you have, the more off-axis frequencies you can cancel, the more directional the system."

I would think it would attenuate the incoming signals.

Yep. I'm thinking in terms of taking a mechanical nightmare and correcting the side effects with multiple microphones and DSP. To be honest, I'm fairly sure it will not be worth the cost or effort as an "interference tube" (single tube with slots) is cheaper, easier, and probably better.

The first one I made was built from the cardboard tubes that the nylon cloth rolls used in my fathers lingerie factory. Visualize a 72" or

This was during the late 1960's, when college protests were in fashion. Both of my microphone versions tended to attract unwanted attention from both sides of any confrontation. One alarmist shouted that the police had a Gattling gun, which immediately cleared about

Psycho-drivel: Using the microphone to pickup music at a distance was a futile effort. The quality just wasn't there. However, for a single public speaker on stage, it worked well both indoors and outdoors. Unfortunately, there was an unanticipated problem. Professional public speakers were so addicted to talking into a microphone that when we took away the mic and used the shotgun mic at a distance, they fumbled awkwardly and demonstrated what appeared to be stage fright. Even a fake microphone on the podium didn't help. Inexperienced speakers didn't have the problem. I theorized that the microphone formed some kind of protective barrier between the speaker and audience, but never pursued the matter.

My trick is to do an image search for "shotgun microphone" and pick out the photos that looks relevant to the type of microphone. In addition to the one on Jim Thomson's web pile, the above two were all I could find.

(

Bell Labs said 300Hz to 3000Hz was required for intelligible conversation. So maybe 16 tubes. But what does the cutoff frequency mean? 3db, 6db, 30db. The highest frequency tube has a Q of 2900hz/450Hz=6.44 Would we need to adjust Q of each tube to match the cutoff frequency, Whatever Cutoff frequency means.

Thanks, Mikek

How about a flat sheet covered with electret mikes, ADCs, and a bunch of DSP?

You could record all the data and do all sorts of analysis, and directional stuff, after the fact.

... but are easy to see if you have one to study. E.g.

Regards, Allan

Combining multiple microphones raises the noise floor. That is one of the design problems in large mixing consoles. It was also a stumbling block for bi-directional CATV systems, until they went to fiber enhanced return channels. Prior to that, it required a complex system of remotely operated relays to select the route to a remote feed. Return amps were only inserted into trunk amps where you needed them. Otherwise, it was impossible to get a clean feed on a medium or large system.

Look at antenna combiners. Two on the same frequency give you a 3 dB gain. Without extensive filtering, multiple mics would overlap and that would add noise.

And charge consumers worldwide a premium? :)

Put two microphones in series and feed them with the same audio. The _voltage_ will be doubled (+6 dB), while the non-coherent noise _power_ will also be doubled (+3 dB), Thus, there is a 3 dB net advantage in SNR.

Combining microphones carrying the same signal raises the random noise by 3dB and the coherent signal by 6dB, giving a 3dB improvement in S/N ratio. The degredation in mixing consoles is because the microphones are not carrying coherent signals.

Only if both are picking up the same exact signal. The whole idea is a group of different frequency bands, in parallel. Also, Electrets need DC to power them. How many isolated, low noise power supplies do you want to build?