Conventional sonar only has one "beam", which is generally not very focussed and gives a mean depth of a larger area of seabed than the individual transducers in a multibeam do. The array of beams in the typical multibeam system gives not only a more accurate, higher resolution picture of the bottom immediately under the ship, they also covers the sea floor to the sides of the ship too. You can perhaps think of conventional sonar as taking very fuzzy wide angle pictures of the seafloor, whilst multibeam builds up a very sharp picture of the sea floor from collating the information from a series of cameras equipped with many telephoto lenses that focus sharply on small areas of sea bed.

This technique works by firing several pin-point acoustic (sound) beams at the seafloor. Each beam is made at a narrowly focussed tranducer which sends out short pulses of sound, and times how long it takes for that "ping" to return to the transducer. With the known speed of transmission of sound through the water and the time this echo takes, one can calculate exactly how far away from the transducer the bottom is.

The system can be so accurate that we have to take into account the movements of the ship in order to make sure we're building up the picture of the seabed correctly despite the heading, pitch (back and forwards like a see-saw), roll (side to side in the typical sea-sickness inducing motion of boats) and yaw (left and right movements) of the ship on the surface, which is done by sensors which capture the motion of the ship in great detail and allow these motions to be taken into account during the processing of the data.

In order to map this data onto the real world, very accurate GPS receivers are used to track the motion of the ship.

In 2002, we used a Reson SeaBat 8111 system to map the canyons of Sodwana Bay.