Voice beamforming/steering mic array

Discussion in 'The Projects Forum' started by emceedub, May 29, 2014.

  1. emceedub

    Thread Starter New Member

    May 29, 2014
    7
    0
    I am interested in building a basic beamforming circuit to explore the concept. I would like to be able to change the signal acquisition direction through adjusting mic output delays.

    I'm sure there are somewhat comprehensive sources on this subject (and have searched the posting archives here for beamforming), but I haven't found them.

    Some questions:
    Are there any basic mic array circuits and/or documented builds suitable for this purpose that the community could point me toward?
    Would filtering the mics to a typical voice 30-3400hz band improve or degrade system performance?
    How many mics would be necessary to get a minimum degree of performance?
     
  2. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,348
  3. sirch2

    Well-Known Member

    Jan 21, 2013
    1,008
    351
  4. emceedub

    Thread Starter New Member

    May 29, 2014
    7
    0
    Thanks for those links, they are helpful background on mics and I will keep them at hand.

    I'm really trying to learn more specifically about arrays and beamforming implementation.
     
  5. emceedub

    Thread Starter New Member

    May 29, 2014
    7
    0
    Thank you for the pointers Sirch2...and a very interesting project.

    I'll try to incorporate the transmission side of things into my thought process. I'm not sure how yours would necessarily be simpler than what I'm trying to do: at this point, I imagine I will just start with 3 mics and see where I get. I'll take a look through the refs at the wikipedia article. So far the lab site (http://www.labbookpages.co.uk/audio/beamforming/delaySum.html) looks most useful for implementation advice.

     
    Last edited: May 30, 2014
  6. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,348
  7. emceedub

    Thread Starter New Member

    May 29, 2014
    7
    0
    Hi Bertus,

    A parabolic reflector would work with a single fixed direction. However, I'm interested in handling varying directions and sources in a stationary fashion. From my understanding, I think a set of variable delay functions would be best suited, though I'm certainly open to other options. Particularly as I'm not sure yet how I'd implement the variable delays.

     
  8. bertus

    Administrator

    Apr 5, 2008
    15,648
    2,348
  9. sirch2

    Well-Known Member

    Jan 21, 2013
    1,008
    351
    I don't think 3 mics is going to do very much because in one dimension you effectively only have one mic. The seven in my arrangement was about a few as is sensible and got the beam angle down to about 11° IIRC. You woiuld be better off with a grid of 9.

    I found a really good on-line calculator that would compute shape, beam angle, etc. but I can't find it now
     
  10. emceedub

    Thread Starter New Member

    May 29, 2014
    7
    0
    Thanks for that insight Sirch2.

    I'm looking at an implementation (essentially a conference table or other circle of sources) with sources in the range of 1-4 meters spread around 360 degrees. So, I wouldn't be assuming a parallel wavefront, and was thinking about a triangular mic arrangement, which would actually exploit that a little. Do you think even at proof of concept level that wouldn't work? I guess to get to 9 I need to pass through 3...so, might as well start and see.

    I'll google around for the calculator you mention.


     
Loading...