Can a horn work backward?

[Chuck]

I’m working with a group on a design that gathers sound from an excitation and analyzes it for the purpose of evaluating the material on which the sound is made. Like knocking on a wall to see where the stud is, only more to it.

Some of the people have been sold on using a PA horn to collect the data, putting the collecting microphone at the throat of the horn. The reasoning is that because the horn is heavily weighted in the audio mid-range - which is the range we’re most interested in - that it will work in reverse: reject the high frequencies and the low frequencies and do a great job of collecting the sound for the microphone. The sound is being generated six inches below the horn. However, there was concern that the sound getting funneled into the horn might distort it and maybe create artificial peaks and valleys. Then the software would be analyzing compromised data. Does anyone have an opinion on that? Thank you for your time!
Chuck

[SeisTres]

I’m working with a group on a design that gathers sound from an excitation and analyzes it for the purpose of evaluating the material on which the sound is made. Like knocking on a wall to see where the stud is, only more to it.

Some of the people have been sold on using a PA horn to collect the data, putting the collecting microphone at the throat of the horn. The reasoning is that because the horn is heavily weighted in the audio mid-range - which is the range we’re most interested in - that it will work in reverse: reject the high frequencies and the low frequencies and do a great job of collecting the sound for the microphone. The sound is being generated six inches below the horn. However, there was concern that the sound getting funneled into the horn might distort it and maybe create artificial peaks and valleys. Then the software would be analyzing compromised data. Does anyone have an opinion on that? Thank you for your time!
Chuck

Yes, this is how the really really old hearing aids worked. They were small backward trumpet for the ear.

And also yes, to your second question, it will distort the source creating unnecessary problems. Best way to measure something is to to NOT use a regular microphone as they have a peaks and dips to make them sound nice, but use a measurement microphone. Then record exactly as it is without any alterations.

[Chuck]

Thank you for your quick response! Is there a consensus from the board on those thoughts? Anyone else?

[Gregory East]

Gut says the measurement mic plan wins unless there is some problem with level of extraneous noise overpowering the mic. Acoustic engineers are thin on the ground here. We build horns but don't necessarily understand the finer points.

[Chuck]

Great post - thank you for your honesty!

[Grant Bunter]

Of course you could always correlate the differences in measurement.
Start with a trumpet (you'd need to hold down the valves) or, since it has no valves, perhaps even better, a bugle, they are both horns.
Use your measurement mike at the mouthpiece end after pointing the bell at the source to get one reading, and compare that to measurement mike straight at source by itself for another, to see if there is any differences.
And see if the two measurements extrapolate to proving/disproving your theory...