microphone s/n improvement

[2dye4]

Hello

I need to improve the sound collecting ability in a
location where there is little background sound energy.
Can I connect several microphones in parrallel to improve
the signal to noise ratio. In other words I want to use
several cheap microphones instead of one expensive one.

Thanks in advance for help

2dye4

 

[felixc]

Multiple microphones will also sum the background noise.

My feeling is that you will get the best improvement with a more sensitive microphone. You may also want one with a cardioid response pattern. Can you also mechanically filter the sound, I mean adding reflectors to improve the gain in the desired direction, or to diminish the unwanted sources, like these microphones with parabolic reflectors used in the football games to pick up the voice of the officials.

 

[2dye4]

I actually am looking to improve the weak signal
sensitivity. I refer to noise as the electrical noise
generated by a microphone without sound present.
If several were paralleled this noise would cancel
but the sound signals are corellated and would add.
Is this possible ??

thanks
2dye4

 

[IRstuff]

You're assuming that all the noise is coming from the electronics. Unless you are doing this in an anechoic chamber, there are few instances where that's really the case. Ambient noise is quite prevalent.

Bear in mind that you would need a minimum of 4 microphones to get a 6 dB improvement in signal to noise.

If your electronics is that noisy, I could get better electronics to start with. Particularly if you're talking about 2 or 4 times the number of microphones and amplifiers. You are almost always better off starting with a cleaner signal than to try an clean up a signal after the fact. You could easily get sufficient phase error in the positioning and equipment to mess up whatever you're trying to measure.

As a final note, a decent configuration and electronics allows the Army's BAT submunition to detect tank acoustic signals from kilometers away.

TTFN

 

[felixc]

Electrical noise per se is random signal. Being random signal it would not cancel the random signal produced by another device but just add to it, in the same proportion that the wanted signal will also add. At the end you won't find any improvement in the signal to noise ratio.

 

[2dye4]

Thanks to all for your input.

Some clarification.
What i would like to do is use cheap pc mount microphones
that cost around $.60 each. They are physically small
enough to mount ten of them together on a pc board in
a 2" by 2" square. With this close proximity the sound
recieved should be in phase. Each microphone feeds through
a resistor to the inverting input of a operational
amplifier configured as a summing amp. So ten mics and
resistors and 1 low noise op-amp. I think the electrical
noise from each mic will be gausian and uncorrelated with
each other while the input sound signals will be
correlated. The noise will add by factor of sqr(2).
My goal is get the sensitivity up much closer to the
human ear.
Does the noise in this case add by (sqr(2))^10 or is this
equation not correct.
Thanks
2dye4

 

[IRstuff]

Weel, you get what you pay for. Ten mics plus amps comes out to over $6 in parts. I'm not convinced that you'd get any better performance than even a $6 mic.


Additionally, you've created a bigger antenna for picking up stray noise and a bigger collecting aperture for picking up ambient noise.

TTFN

 

[nbucska]

If you increase the distance, you will get better directional effect for lower frequencies, too.

Write a computer model !


<nbucska@pcperipherals.com>

 

[felixc]

Mounting them as an array will effectively create a frontal lobe where the sensitivity is better. The equivalent of a physical reflector. This will improve your S/N ratio in this direction, more than the sum of the noises of each microphone. This is the principle of sonar arrays, although a 3x3 array will not give a hell of an improvement, but it will make a difference. Does it make the array of cheap transducers better than a single higher-quality microphone, the only way to know is to try it.

 

[circuit1952]

2dye4,

IRstuff and felixc have some valid reasons why what you suggest is a bad idea. If you start with a good quality microphone, connectors, cable and electronics, you are better off than trying to clean up the signal.

You have said that you want to achieve the sensitivity of the human ear with an array of really cheap microphones and you assume that the noise from said microphones is electrical rather than acoustic. How about some more details:
Is this to be used indoors or outdoors? Would a windscreen help?
Is the source fixed in relation to the microphone (in direction if not in distance)? Can a directional mic be used? If the direction isn't fixed, can a directional microphone be rotated toward the source?
What frequency range do you need to cover? At some frequencies and some microphone spacings the signal will cancel!

 

[2dye4]

Thanks all for your help.

This will be indoor use.
I understand a reflector for sound has to be some
significant portion of a wavelength for the lowest
frequency of interest. If this is true it makes the
dimension too large.

All electronics are on the same pc board and a
high quality low noise op-amp will be used for the
summing and preamp.
thanks

 

[IRstuff]

Getting rid of noise by post-processing is almost always a poor starting point. Any processing that removes the noise will remove some signal and corrupt the signal in some fashion.

TTFN

 

[nbucska]

It will take a lot of work to approach a good microphone.
If your time is money, it is a loosing proposition.

<nbucska@pcperipherals.com>