Fishes have evolved multiple mechanisms for sound production, many of which utilize sonic muscles that vibrate the swimbladder or the rubbing of bony elements. Sonic muscles are among the fastest muscles in vertebrates and typically drive the swimbladder to produce one sound cycle per contraction. These muscles may be extrinsic, typically extending from the head to the swimbladder, or intrinsic, likely a more-derived condition, in which muscles attach exclusively to the bladder wall. Recently discovered in Ophidiiform fishes, slow muscles stretch the swimbladder and associated tendons, allowing sound production by rebound (cock and release). In glaucosomatids, fast muscles produce a weak sound followed by a louder one, again produced by rebound, which may reflect an intermediate in the evolution of slow to superfast sonic muscles. Historically, the swimbladder has been modeled as an underwater resonant bubble. We provide evidence for an alternative hypothesis, namely that bladder sounds are driven as a forced rather than a resonant response, thus accounting for broad tuning, rapid damping, and directionality of fish sounds. Cases of sounds that damp slowly, an indication of resonance, are associated with tendons or bones that continue to vibrate and hence drive multiple cycles of swimbladder sound. Stridulation sounds, best studied in catfishes and damselfishes, are produced, respectively, as a series of quick jerks causing rubbing of a ribbed process against a rough surface or rapid jaw closing mediated by a specialized tendon. A cladogram of sonic fishes suggests that fish sound production has arisen independently multiple times.