It's a commonly known fact that whale song travels huge distances. In fact apparently before the oceans were so noisy they could communicate over 10,000km!

However, whales are very big! They can create very loud noises, far louder than any being of 2 metres. In addition whalesong is typically around 10-40hz and is lower than a human is capable of producing. I'm unsure if this has bandwidth implications.


  • A creature of about two meters in length
  • Wanting to hold a normal conversation spoken (in their own language but basically humanlike speech)
  • Assuming they can't open their mouths throat type sounds are permitted/encouraged
  • The creature must also be able to communicate with humans (speak a human language as a second language)

What is a reasonable range they could "shout" to their friends underwater? 100 metres? 400? A mile!?

  • 1
    $\begingroup$ As a starting point: the world record for lowest note sung by a human is a D2, which corresponds to 73.416Hz. $\endgroup$
    – overactor
    Commented Oct 14, 2014 at 13:03
  • 2
    $\begingroup$ Someone please answer with a detailed discussion of the (from what I understand) stupidly-immense distances that whale-song can carry over. Especially if the constraints of the answer affect it (I expect the size limit to be relevant). $\endgroup$
    – KRyan
    Commented Oct 20, 2014 at 15:08
  • $\begingroup$ @KRyan I'd love that to be the answer - imagine a species which could communicate over thousands of miles in a language completely impervious to outsiders... it would be awesome! $\endgroup$
    – Liath
    Commented Oct 20, 2014 at 15:11
  • $\begingroup$ I find human voice box really limiting here...dolphins lack the voice box in our manner, but they can produce sounds in the 7-15 khz range that should travel quite a distance. Human voice box doesn't seem to perform well in water $\endgroup$
    – Twelfth
    Commented Oct 20, 2014 at 20:17
  • 1
    $\begingroup$ @liath - I'm a little conflicted then...an Aquatic creature that evolved speech designed to travel across airways doesn't make the most sense. A hybrid creature could possibly develop two communication methods in conjunction (one designed as a 'song' through the water ways similar to a dolphin (whistle and clicking), and a second 'humanesque' air communication using lips tongue and teeth). Interested in a hybrid species, or are we talking only within the realms of human speech? $\endgroup$
    – Twelfth
    Commented Oct 21, 2014 at 19:09

6 Answers 6


TL;DR: Normal conversation at 1 km depth would travel about 2 kilometers in water. The loudest scream a human could produce would travel about 3.14 km.

There are many factors that make it difficult to calculate how far sound travels, the biggest being pressure and temperature. The lower the temperature, the less far sound travels, the higher the pressure the farther it goes.

Calculation of the distance sound goes is difficult because not all the variables are known. But the principle behind sound losing volume over distance is called absorption. The sound also loses energy as it spreads out. The total energy loss of these two factors is called Transmission loss. Depending on the chemicals in the water, different amounts of sound are lost over different distances.

To do some math trying to get average values. The unit for sound energy is dB. Normal conversation has a power of 60 dB. In water, sound is stronger, so we add 62 to all values for sound in the unit dB. This means that normal conversation under water has the effective strength of about 122 dB. Normal voice frequency varies, but a typical frequency would be around 170 Hz. At about that frequency in most oceans, the absorption rate is about 0.06 dB/km. So we can see that the absorption rate is negligible (Though it will be included in the below calculations). But this does mean that frequencies have a small effect on the distance sound travels.

Now, sound also loses energy as it spreads out. Sounds spreads out in two different ways, spherically, and cylindrical. When the water is deep enough, it spreads out spherically, but when the water is more shallow, it spreads out cylindrically. This happens because the sound bounces off the surface of the water and the bottom of the ocean. So while some energy is lost when the sound bounces, more energy is lost when it travels out fully in a sphere. So sound travels approximately twice as far when using the cylinderical form. Sound spreads out in a cylinder when the sound does not fade before it reaches the top of the water (so it bounces off the surface) The amount by which the sounds distance dilutes over a given distance can be calculated. The formula for the loss here is TL = 20 Log(R). TL stands for Transmission Loss, and R stands for range/radius of the sphere. Because dB is logarithmic scale, we can directly transfer this into dB units. So at 1 km, this comes out to 60 dB, which perfectly matches the volume of human conversation.

So the total Transmission Loss is 60 dB/km + the absorption rate, which is 0.06 dB/km, which comes out to 60.06 dB/km

Generally, the threshold of human hearing is 0 dB. So using the data calculated above, the sound of normal human conversation (if it could somehow be communicated at the same strength underwater), would travel about 2.03 km.

Wikipedia says that the loudest recorded scream was 129 dB, which is 191 dB in water. So using the above calculations, at 1 km depth, a scream would travel about 3.18 km. Note, that is high enough volume to hurt human ears, so the underwater creatures should be able to handle louder volumes, or they should speak quieter.

There is one phenomenon that lets sound travel for much longer distances. This has been especially noticed from the noises whales make and it allows them to communicate over very large distances. When sound travels through water it slows down, causing it to refract downward. The water below keeps getting cooler, so the sound keeps slowing down and refracting. At a certain point, the water stops getting colder, but the pressure continues to increase. When pressure gets higher, sound speeds up. This causes the sound to refract upward. It then refracts back down as it gets back into areas where the water temperature slows. This refracting up and down allows the sound to travel a very long distance without losing much energy.


So there are a lot of factors that affect the distance sound travels, the biggest being pressure. At the standard pressure used by scientists the sound of normal human conversation would travel about 2km. But at certain depths, sound actually travels for a very long distance. This phenomenon is what allows whales to communicate over thousands of km. To a lesser extent it could work for human voices too, if not over crowded.

And finally I would like to say that this data is not fool-proof. There is no easy way to calculate exact values, so these values are not exact.

References and Notes:

There are some very good graphs showing absorption rates on this website. I got sound values and conversion rates from air dB to water dB from this site. Unfortunately, copyright prevents me from re-posting them here. For information on the full math I did to get the base TL for spreading out, see this site. Voice frequency information came from Wikipedia, as did scream information. Thanks to Irigi for pointing out some mistakes in my first drafts.

  • $\begingroup$ -1: - it seems to me the answer is wrong. It is nicely formulated, but the facts do not match. Absorption rate of seawater at frequency ~100 Hz seem to be around 10^-3 dB/km. Moreover, you are only calculating with absorption, but you do not take into account the fact that the sound gets weaker as is travels to all directions. $\endgroup$
    – Irigi
    Commented Oct 21, 2014 at 10:12
  • 1
    $\begingroup$ @Irigi I would like to point out that the absorption rate is actually closer to 10^1 db/KM, and I'm going at 170 Hz, so we're looking at different values. I'll look into what happens when sound spreads out. $\endgroup$
    – DonyorM
    Commented Oct 21, 2014 at 10:46
  • $\begingroup$ Sorry, 10^-3 dB/km is my mistake. One should look at the values between f 10^0 and 10^1 in the graph in the reference. (100 Hz - 1000 Hz). So the correct value is closer to 0.1-1 dB/km. But 10-100 dB/km would require 5 kHz-200 kHz, which is not common in speach. Still, sounds around 10 Hz would have absorption around 10^-3 dB/km, and such could probably be produced by whale-like creatures. $\endgroup$
    – Irigi
    Commented Oct 21, 2014 at 11:10
  • 1
    $\begingroup$ @Irigi finally, I understand the chart, thanks for working that through with me (despite how dense I was). I think this is finally fixed, though I will check again when I have time. $\endgroup$
    – DonyorM
    Commented Oct 21, 2014 at 12:37
  • 1
    $\begingroup$ @Irigi, I did vaguely include this, but I may aswell make it clearer. $\endgroup$
    – DonyorM
    Commented Oct 21, 2014 at 13:15

Might be way off with this, but what the hey.

I don't think you'll get very 'human like' speech out of a race that developed language under water. Humans fit into a weird category where our hearing 'system' evolved from aquatic creatures (same deal with our sight), quite likely from early fish...however our language itself developed in more modern times when mammals freely roamed the land. So our vocals became directly tied to the medium (air) in which we spoke through and not the water our hearing stems from.

A species that remained underwater would not have that 'vocals tied to air' attribute to them and would quite likely develop speech directly tied to the water as a medium. Are you aware of 'tonal' languages? It's a foreign concept for a native English speaker (except for a few rare examples...an upwards inflection at the end of a sentence will denote a possible question for example), but to people used to detecting it, these tones mean as much as the letters. Mandarin contains the 'ma' case (it's the reverse of the rice/lice difficulties they have)...the word 'ma' has 4 different meanings... “mother”, “to scold”, “horse”, “hemp”. The difference between these 4 terms is entirely tonal..a falling tone vs a rising tone vs a high tone is the difference in these words (the end result here is english speakers saying they are off to feed their mother hay and kiss the horse goodnight).

In the case of an underwater species, there are several sounds in English that would be pretty much useless. Mouth out the 'p' sound and then the 'f' sound...both these sounds are a puff of air with lips touching for p and not touching for f (Indonesian languages lack this distinction). Same goes with the r vs l (say earl fuller to get the differences in tongue position there)...th vs d is simply a tongue flick off the top of the roof of your mouth vs the back of your teeth (a tongue motion the french lack resulting in them saying dis and den. English lack the French OE sound, which is kinda like saying oooo while your mouth is in the smiling eee shape).

This limitation on 'mouthed' sounds not being effective underwater would do much to limit the language in terms of sounds that could be effectively communicated. This gives the tonal nature of an underwater language that much more feasible as a fill in for letters that they can't quite do in an underwater sense (would also produce languages that sounded a little more like a whales song than a human language...almost a whistling language where the changing pitch of a letter means as much to the meaning of the word they are saying as the word itself).

This is a very long way of saying that the English language underwater only has an effective range of a good 5-10 feet before things start to get lost, and the majority of the sounds we are dependent on in determination of words are near impossible to effectively communicate. An aquatic species that developed on it's own would much more likely have a 'song' language of their own that could have a communication range of well over a few KM had the intentionally made a sound for long distance travel (IE a shout)

  • $\begingroup$ "Humans fit into a weird category where our hearing 'system' evolved from aquatic creatures (same deal with our sight), quite likely from early fish" Citation for this claim? $\endgroup$
    – ubadub
    Commented Oct 3, 2018 at 4:22

In general low frequencies travel further but carry less information in a given time span. As a result you have a trade-off between those two competing requirements. In addition low frequency noise generally needs the creature making the noise to be larger (although there are ways around that to a certain extent with specially adapted organs).

Water transmits sound much more effectively than air. In air the speed of sound is 342 m/s, in water it is 1,484 m/s. In terms of loss of volume there is very little absorption of low frequency sound (one reason it carries so long) but it is still spreading out and getting quieter as it does so. The water-air barrier acts as an almost perfect reflector so that does tend to reduce some of the loss but can create echoes and similar confusions.

This wikipedia article may be helpful for you, it covers the subject of Underwater Acoustics. In particular there are several sections on the propagation of sound underwater.

  • 2
    $\begingroup$ there's also temperature inversion that will provide another reflection surface, (though less perfect) $\endgroup$ Commented Oct 14, 2014 at 15:58

Dolphin echolocation works at a max range of about 200m. That is at a slightly higher frequency though(40 to 130 kHz). Also that is 200m round trip so 400m one way, and accurate enough to target something the size of a couple of inches. For general singing or shouting I think something in the neighborhood of 1km sounds reasonable.


Absorption of sound is actually a very minimal thing. The bigger issue is range losses. Sound power drops off by range^2 as it propagates outwards, and that limits most sound. The low frequencies of whale communications DO help (frequency does matter), but the more important factor for whale communication is that they are leveraging the Deep Sea Channel, which abuses pressure gradients to create a "wave guide" that keeps the sound from going up or down. This decreases those propagation costs to merely range rather than range^2.

A bigger limitation is your requirement that they hold a "normal" conversation. Any acoustic environment can be modeled as a noisy communication channel, where "signal to noise ratio" matters. Even whales have limited communication in areas where the sound of the surf is too noisy. As well, the higher the information content, the higher quality the channel has to be.

Consider making communication more directional. If you think about it, a "shout" across the ocean is heard by all. If there's thousands and thousands of creatures all "shouting" at once, they'll drown each other out.


It might be plausible that an undersea creature with a fully optimized low frequency, long distance, like to like, communication system as a whale, might evolve a completely separate surface communication system. Perhaps a dolphin like blowhole that served some symbiotic purpose like calling birds to clear an obstructed hole, or was used like a game call to lure land animals close to the waters edge to be devoured. The low frequency system could be non vocal, triggered by whole body shuddering. The higher frequency air communication system while initially evolved as biomimicry, could further be naturally or artificially selected as sick animals that made sounds humans understood were more likely to be given antibiotics, or other treatment.

Artificial selection is when humans choose to breed animals with specific traits that are pleasing to them, despite being potentially damaging to the animal. ie. fainting goats, dogs with snouts so short they cannot breathe properly, great danes and St. Bernards so big they have heart trouble. It seems such a process might need to play a large role here.

  • $\begingroup$ While this is interesting, it doesn't seem to answer the question, which relates to the range at which they could communicate. $\endgroup$
    – Brythan
    Commented Jun 20, 2016 at 23:34
  • $\begingroup$ The OP already knows the range of whales, but seemed to think that range needed to be reduced by some factor due to "assuming they can't open their mouths" and "able to communicate with humans" I was trying to show that no compromise needs to be made. $\endgroup$
    – slomobile
    Commented Jun 21, 2016 at 1:48

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .