Best Of

Honestly, most of it sounds like a nail scratching glass: a sharp, intermittent, pitch-variable sound. The frequency range depends very strongly on what you're trying to do; mapping the seafloor generally uses mid- to high-pitch, while the most effective pitches for finding very quiet subs are low. (The US Navy actually deployed a low-frequency sweep system for detecting superquiet subs in the late 90's, but the tones were found to be harmful to marine life, so the Navy agreed to limit its use.) Ping structure varies depending on what you're doing, and some pings will (at least initially) sound like just a chirp, but if you listen to the reflections you can tell that it's actually a sweep.

When you send a sweep, you get a very distorted sweep back from a reflective object. By mapping those distortions, you can determine a lot of things: distance, relative size, and if it's mobile, a general idea of its speed and heading. (It's also a damn ghostly sound.) But when you send a ping, all you get back is, well, a ping. Then you need to send a whole bunch more, steering the sonar head around and listening to each one to try to find what on Earth is bouncing your ping back to you and from where. The old ASDIC system worked that way, as did the most primitive SONAR rigs - WWII sonar was exclusively the steerable transducer type. All of that was thrown out the moment more precise and data-dense sonar became available.