14

I'd like to estimate the distance at which the bird is singing from the point of recording, to then finally get the source level from one microphone.

I've seen studies, and heard from colleagues use visual guesstimates and laser range finders. Guesstimation is good, though errors when the bird is nearby (a distance of 1 m instead of 0.5 m makes a big difference to the estimated source level). The laser range finder I have at least doesn't really work well during the day (it's also not a low-end model...Bosch GLM 50).

What are some of the ways I can estimate distance to the bird accurately (or any other diurnal vocalising animal) during daylight?

Post-script: The question was initially ambiguous about the number of microphones in play. It has now been edited to specify that Nmics=1.

5
  • I think the title of your question doesn't match your question (distance vs source level) Jul 1 at 9:52
  • 1
    Distance is needed to calculate SL, so I'd argue they do match
    – Chloe
    Jul 1 at 12:51
  • I've added distance to the question just in case the SL-distance connection isn't clear for someone searching in a hurry.
    – Thejasvi
    Jul 3 at 7:38
  • based on variation in answers, it would be helpful to specify whether your recordings were made with 1 device or with an array of time-synced devices
    – Chloe
    Jul 4 at 14:36
  • 1
    Done. Added one mic to the question.
    – Thejasvi
    Jul 4 at 14:48

6 Answers 6

6

Would it be possible to use a video camera in the field, and make video recording concurrent with audio recordings? You could then take note of the locations of birds on different branches/locations whenever there is an audio detection, and then manually measure that distance to the recorder at a later time (or use your lazer range finder at night instead to get the range to the location)? I know this is not perfect and assumes that you know, visually, which bird was vocalizing (and assumes you can see the birds). Another approach could be to use a drone with an altimeter to get the height of a bird location, and then use the Pythagorean theorum to get the distance to your recorder.

1
  • (Answered assuming only a single microphone made measurements and so acoustic localisation would be unavailable)
    – Chloe
    Jul 5 at 14:36
5

Acoustically localising angles is usually far easier than localising distances (hardware and analysis is greatly simplified). You could use 4 closely spaced synchronized microphones (e.g. 10cm apart) in a tetrahedral configuration attached to a video recorder (similar idea in underwater acoustics here). The microphones will give you the relative horizontal and vertical angle of a vocalising animal in the video recording; you then know which animal is making a particular sound and have a have a few options for distance.

  1. As @chloe says - if you know the height of the bird then distance is easy using simple geometry. Note that you know the height of the sea surface so it's easy to get the distance for (non-flying) marine animals. Also, if you are interested in ranges <5m then you can reconstruct 3D environments using LidDAR on an iPhone or iPad (www.apple.com). Other commercial LiDAR systems have greater range.

  2. You could use another tetrahedral array some distance away to get another set of angles/vectors to a vocalizing animal - the intersection of two (or more) vectors is the location of the animal (we did something like this here for porpoises). Note that is very important to know the exact position and orientation of each tetrahedral microphone array for this to work.

  3. Use some sort of photogrammetry method (e.g. two cameras) to determine location.

3

Acoustic localisation or using a total station is neat, of course, but in many cases (at least for my work) a laser range finder and a good old tape measure for distances below 10 meters will do the trick. I think the Bosch laser range finders are mainly designed for indoor use. I've never had any problems taking distance measures in broad daylight with a Leica range finder (Leica Rangemaster 800).

1
  • Awesome point about indoor vs outdoor laser range finders! Do you mind sharing the model or series in an edited version of the answer?
    – Thejasvi
    Jul 4 at 10:56
3

I think most of the answers so far are suggesting you localize via acoustics, while your question seemed to be hinting at measuring distances without the acoustics. If the idea is to try to evaluate animal source levels it seems to me that having a "real life measurment" that does not require the acoustic data to do so is probably best - this is based on the motto that "if you can measure something do so, if not model and estimate it" (I might have just made up that motto). Depending on the scenario, making an exact record of the position of the animal (e.g. taking a bunch of digital photos of the animal while it is making sounds) and then measuring with a tape measure afterwards the exact distance to the cm to your sensor seems a reasonable (and possible optimal) approach to me.

2

The best way to do this is localization. It won't be easy though. You will need at least three acoustic sensors placed well apart. Search the acoustical society of America for 'Lombard effect' and a lot of the papers will discuss how they estimated source levels of calling animals.

0

You could use a Total Station to estimate the xyz coordinates and distance of the bird at hand.

Of course a Total Station is a slightly more financially and logistically involved solution (low-end models start at >= 1000 Eu), and is another piece of bulky equipment to carry around and assumes relatively steady ground for the tripod!

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.