Is it possible to understand which way sound source is located with using microphones? I was thinking using multiple mikes pointed out different ways and getting db differences to predict direction. Would it be accurate?

Edit: Lets say we have control of sound source. For example I can play a instant buzz at a specific frequency from source. Can I use time delays between microphones for getting direction. Is that process can be performed real-time. And if it can possible what kind of hardware should I use.

  • \$\begingroup\$ How many ways can a sound source be located? \$\endgroup\$
    – Andy aka
    Dec 9 '15 at 18:01
  • \$\begingroup\$ Would it be accurate? Depends. Will your operation site match the assumptions you make while deriving your predictor? \$\endgroup\$
    – MickLH
    Dec 9 '15 at 18:06
  • 2
    \$\begingroup\$ It would be better to triangulate based on time delays. If you have three or 4 microphones, you can tell where the sound is coming from by measuring the relative time delays to the microphones. The one that heard it first is closest to the source, etc. Using trigonometry you can calculate the exact angle (in 2D). \$\endgroup\$
    – mkeith
    Dec 9 '15 at 18:40
  • 1
    \$\begingroup\$ better to use omnidirectional microphones and measure the time delay. example code for measuring distance between microphones: gist.github.com/endolith/376572 or combine amplitude and delay information, like the DUET algorithm \$\endgroup\$
    – endolith
    Dec 9 '15 at 19:19

Yes, the most common way is with phase correlation, which takes some amount of processing power. Basically you have two (or more) microphones a known distance apart and measure the time difference you see between the same sound in the different microphones.

IMO using amplitude differences would be very inaccurate.


While it is possible, your approach generally won't work. The problem is that, for reasonable systems, the difference in intensity between two microphones is just too low to be useful. An example of the approach working (more or less) was the acoustic mirrors used by Britain prior to WWII. Detection of phase differences can work, but this is generally restricted to detecting time of arrival for impulse noises. Various cities have installed gunfire locators.

  • \$\begingroup\$ You can use correlation of complex voltage as per @Daniel \$\endgroup\$
    – johnnymopo
    Dec 9 '15 at 18:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.