What did dinosaurs sound like?

No human heard a non-avian dinosaur, and sound itself does not fossilise. There is therefore no single defensible recording for Tyrannosaurus, Velociraptor or Parasaurolophus. Film roars mix voices of living animals with processed effects. They can be dramatic without becoming scientific evidence.

Several independent lines still narrow the possibilities. Rare laryngeal elements preserve part of a vocal tract. Hollow crests have calculable resonances. Inner ears constrain frequencies an animal could perceive. Birds and crocodilians provide comparative boundaries because they are the two living archosaur branches. Each line answers only one part of the question.

A voice needs more than a skull

Vocal sound requires a vibrating source, airflow and a tract that filters the resulting spectrum. In most terrestrial vertebrates the larynx at the top of the airway supplies the source. Modern birds mainly generate sound lower down, in the syrinx where the trachea divides into the bronchi. Beak, tongue, pharynx, oesophagus, nasal passages and inflatable soft tissues can then reshape it.

Non-avian dinosaurs usually preserve skull bone, parts of the hyoid apparatus and occasionally tracheal rings. Most cartilage, membranes and muscle disappear. Even an excellent skull cannot show which tissue vibrated, how exhalation was controlled or whether the mouth stood open. Bone limits the options but rarely selects one timbre.

Body size is not a direct musical calculator. Large animals often can make low signals because their airways and resonating spaces are long, yet pitch also depends on the vibrating element. A small membrane can create a high source tone that a long tract modifies. Mass alone cannot yield a dinosaur note.

Why birds and crocodilians matter

Birds are living dinosaurs, while crocodilians represent the other surviving archosaur branch. A feature shared by both groups is more plausibly inherited from their common ancestry. When their mechanisms differ, an extinct dinosaur may fall within a range rather than match either endpoint.

That is exactly the acoustic situation. Crocodilians use a larynx and birds a syrinx, but both can hiss, call with an open mouth and produce low closed-mouth signals. Comparison supports diverse acoustic communication among dinosaurs. It does not turn an alligator bellow or cassowary call into the voice of a particular fossil species.

The evolutionary context is covered in early birds and the origin of flight. Relationships provide testable analogues, not a copy of behaviour ready to paste onto every branch.

The Pinacosaurus larynx is rare direct evidence

Specimen IGM 100/3186 of the ankylosaur Pinacosaurus grangeri preserves parts of the hyoid and laryngeal apparatus. Researchers identified a cricoid element, paired arytenoid elements and associated bones. It is the oldest highly informative fossil larynx yet described and the first reported for a non-avian dinosaur.

Its basic inventory resembles other reptile larynges. At the same time, the large cricoid, long arytenoids and mobile joint recall features of birds, in which the larynx reshapes the vocal tract. The authors proposed that Pinacosaurus could modify sound flexibly and may have used bird-like vocal tract control.

Preserved relationships among the pieces are crucial. One small isolated element could be misidentified. In this specimen the structures lie in the throat region and form an anatomically consistent complex. Comparisons with birds, crocodilians and lizards test their identity and possible joint movement.

This fossil does not show that Pinacosaurus sang like a modern bird. Membranes, muscles and the complete soft tract are gone, and no syrinx was found. It supports controllable vocal modification, not a recoverable melody.

Did non-avian dinosaurs possess a syrinx?

The oldest described Mesozoic syrinx belongs to Vegavis iaai, a bird that lived in Antarctica about 69 to 66 million years ago. Mineralised rings and bronchial asymmetry show that a developed avian sound source existed near the end of the Cretaceous within the lineage leading towards modern birds. The find also proves that the organ can fossilise under favourable conditions.

No comparable organ has been found in a non-avian dinosaur. Absence is not an absolute disproof because an early syrinx may have mineralised poorly and fossilised rarely. But the Vegavis structure cannot simply be transferred to Tyrannosaurus. Current evidence allows the specialised avian syrinx to have arisen relatively late in bird evolution.

Flight structures, tails, shoulders and body proportions evolved as a mosaic. The vocal system probably did too, with different components appearing at different times rather than as one package.

A closed mouth can still make sound

Some birds boom, coo or hoot with a closed beak while expanding the throat or oesophageal region. These signals tend to emphasise low frequencies. Crocodilians likewise generate low sound without a widely opened mouth. Comparative work shows that closed-mouth vocalisation evolved repeatedly among birds and occurs especially often in large-bodied lineages.

The capacity is plausible for some non-avian dinosaurs, but it remains an inference from living relatives and body size. Not every large dinosaur must have boomed, and the mechanism need not have dominated communication. Different groups may have combined closed-mouth calls, hisses, higher cries and non-vocal noise.

Low frequencies attenuate less over distance and bend around small obstacles better than high frequencies. They could be useful in forests, across floodplains or within a dispersed group. Usefulness does not prove behaviour. A specific taxon first needs compatible anatomy and hearing sensitivity.

Parasaurolophus and its hollow crest

The long crest of Parasaurolophus is formed by skull bones enclosing elongated nasal passages. Air travelled from the nostrils through the crest and back towards the airway. The system could operate as a resonator that changed and strengthened selected components of an existing signal.

Acoustic calculations for lambeosaurine cavities generally predict low-frequency resonances in adults. The crest was not necessarily a self-playing horn. Oscillation still needed to originate at the larynx or elsewhere in the airway, after which the tubes filtered it.

The crest was visually prominent, differed among species and changed during growth. Acoustic and display functions are therefore compatible. Selection could retain a recognisable silhouette and properties of the inner passages at the same time. A single-purpose explanation is not required.

Juveniles and adults may have sounded different

Lambeosaurine crests changed substantially during growth. The very young Parasaurolophus nicknamed Joe had only a short developing crest even though its body had reached roughly one quarter of adult length. A shorter airway should have different resonances, generally shifted towards higher frequencies.

Age-specific sound might help individuals recognise life stages, but that is a functional hypothesis. The fossil directly records juvenile skull geometry; the acoustic consequence follows from cavity dimensions. No preserved observation shows how a herd responded.

Other hollow-crested hadrosaurs need not match Parasaurolophus. Lambeosaurus had a different outer silhouette and chamber arrangement, so its favoured resonances could differ as well.

The inner ear describes reception, not production

CT data can reconstruct the bony labyrinth. Among living archosaurs, dimensions of the cochlear region correlate with aspects of hearing range. Lambeosaurine anatomy is compatible with sensitivity to the low frequencies predicted for hollow crests. Tyrannosaurids also had an elongated cochlea and extensive middle-ear air spaces, suggesting that low-frequency sound mattered to them.

A match between likely signal and hearing sensitivity strengthens the communication hypothesis, but the ear does not disclose the source. An animal could hear footsteps, environmental noise, prey or conspecific calls. Even an estimated optimum frequency says nothing about melody, rhythm or volume.

How an acoustic model is built

Researchers scan a skull and segment rock from bone across many slices. They reconstruct the inner wall of a nasal passage, repairing cracks and missing sections transparently. The digital cavity becomes an acoustic tube or a fluid-dynamics model. Calculation identifies frequencies at which that geometry should resonate strongly.

Results depend on length, diameter, air temperature, opening positions and assumed soft-tissue boundaries. Different plausible parameters produce different spectra. A careful study publishes those assumptions and reports a range rather than one deceptively exact note.

The model may be checked against a physical tube or a second numerical method. Agreement makes the cavity's resonant behaviour more secure. A museum audio file is then produced by sending a chosen artificial source through the calculated filter. The filter can be well constrained while the starting signal is still human-selected.

Even a physically correct resonance is not a full vocalisation. A complete recording would require the source waveform, laryngeal motion, breath pressure, mouth position and behaviour. Those are usually the missing components.

Tyrannosaurus did not necessarily roar

Film sound designers combine mammals, birds, reptiles and processed noise for large predators. In Tyrannosaurus, inner-ear anatomy supports sensitivity to low frequencies, but the vocal source and complete tract are unknown. A low boom is possible. A recognisable lion-like roar is unsupported.

Uncertainty is equally important for Velociraptor. Its endocranial anatomy has been studied, yet it does not prove film clicks, hisses or cries. Small body size and proximity to birds widen the comparison set without selecting one sound.

The strongest reconstruction offers several acoustically and anatomically compatible outcomes. The more confidently an audio track claims one unmistakable voice, the more artistic choice it contains.

Did dinosaurs communicate vocally?

The rare larynx, probable resonators and developed hearing make vocal communication highly likely in at least some groups. Signals could maintain contact, warn of danger, accompany display, distinguish individuals or convey age. Living archosaurs use sound in all these contexts.

A function must still be tested separately. A bonebed does not preserve a chorus and a hollow crest does not demonstrate a mating call. Visual communication may have been equally important. Crests, frills, horns, posture and colour operate at close range without sound, and one structure can combine display with acoustic filtering.

Noise without a vocal organ

Dinosaurs made sound through footsteps, vegetation movement, contact with the ground and jaw closure. Feathered species might have produced mechanical noise with feathers as some birds do, but no direct evidence assigns that behaviour to a particular non-avian dinosaur.

Footstep volume cannot be calculated from mass alone. Speed, gait, substrate softness and foot anatomy all matter. Tail cracks, body drumming and claw clicks must be tested against strength, movement range and wear. Without such evidence they remain scenarios.

What counts as a responsible sound reconstruction?

A responsible project names the specimen and separates four levels. A laryngeal bone or nasal passage is observation. Tract mobility or calculated resonance is anatomical inference. The purpose of a call is a behavioural hypothesis. The final sound file is an artistic demonstration of one selected model.

A useful recording presents alternatives and states which parameters changed. It does not hide missing soft tissue behind an exact result. Frequency, loudness and timbre require particular caution because they depend on parts of the system least likely to survive.

Conclusion

Dinosaurs were almost certainly not mute, but their exact voices remain unknown. The Pinacosaurus larynx offers rare direct evidence for a complex tract. The Vegavis syrinx shows that an advanced avian source existed by the latest Cretaceous, yet it has not been found in non-avian dinosaurs. The Parasaurolophus crest could filter low sound, while several dinosaur ears fit low-frequency perception.

Together these observations make booms, closed-mouth signals, hisses and varied calls plausible for different groups. They do not support one universal roar. The most honest reconstruction is a testable range with visible assumptions, not a recovered soundtrack from the Mesozoic.