I have started creating a language for a primitive alien civilization I am (slowly) making, and my first step was selecting sounds for this languge.
But I was wondering if aliens could make these sounds with their voice, and more precisely what anatomic characteristics would they require to speak (kind of) the way we do? And is it realistic to have only human-like sounds in an extraterestrial language?
The title question is easily answered: Yes, they are.
Even some animal species with a quite different vocal tract (parrots and ravens) are able to imitate human speech sounds good enough to be recognisable. I don't doubt that the already civilised aliens will be able to do something similar. If anything else fails and their civilisation is advanced enough they will use speech synthesis. It is available to humans since the late 18th century (van Kempelen build a completely mechanical speech syntesiser than).
However, their own language used for internal communication will very probably very different. Just like parrots communicate with each other in their native sounds, the aliens will do so, either.
Some overlap is probably reasonable, but not 100%.
To start with non-human species making human-range sounds, you can look at the animal kingdom, from both directions: First, some animals, like parrots, can replicate human voices (albeit at a different pitch) and other animals like cats or dogs are often recorded making sounds that seem similar to human words. Second, most humans can imitate or approximate the 'normal' sounds of these a wide variety of animals. (Of course, larger animals like elephants or whales tend to be stuck at the lower-pitch end of the spectrum, and small animals like mice tend to be at the higher-pitch)
Now, to ignore animals and aliens entirely: Not all human languages share the same phonemes. Think about someone raised speaking Spanish trying to say "victory" - the first part sounds a lot more like "bic" than "vic", because they just don't have that sound in their language. (If you do have a friend with a strong Spanish accent, another word they will probably have trouble with is "warthog")
So, we can see that non-humans can approximate human speech, but even other humans sometimes have trouble managing all the sounds in a language.
Define "human-like sounds"...
Various African languages use click sounds. Most European languages lack the two "TH" sounds from English. English itself lacks the French "J" sound, and English English lacks the Scottish "CH" sound. French and Finnish both differentiate rolled and unrolled "R" sounds. Japanese has one consonant halfway between "R" and "L" which causes many Japanese to be unable to distinguish the two as separate consonants when speaking or listening to European languages. And then we get onto tone languages like Mandarin.
More unusually, there are even whistle languages.
And that's just for formal language. The human voice is capable of some remarkable sounds, as demonstrated by beatboxers purely for musical and rhythmic effect.
Children are born capable of adapting to all these sounds, and all adults continue to be physically able to produce these sounds (albeit sometimes requiring practise, and demonstration from other people of how to produce them). However during early development, babies lose the ability to hear vocal sounds other than those around them, so that by the time they reach the "babbling" stage, the sounds they produce are only those sounds which they have already heard from people around them.
Given the diversity of sounds possible from the human voice, a similarly capable alien language may have a similar diversity of sounds. The problem you have then is explaining why the aliens only have a single language! (The old Star Trek problem of "one village represents the whole planet".)
First is 20-20k Hz a reasonable range for hearing? --Do we likely hear the same sounds they do?
In gases of pretty much whatever type pass all frequencies pretty well. If they live in some kind of air they will not be constrained to our hearing and could as easily use something above it as in it. If they live in some other medium that is better at absorbing energy like water or in the ground, lower frequencies can go much farther, so they might prefer frequencies below our hearing.
Many natural events produce distinctive patterns inside our hearing range (you know what a footstep sounds like), but actually produce more energy in other bands (footsteps peak in the single hertz range).
In general you can move more information at higher frequencies, but send it farther or through worse obstacles at lower frequencies. If maximizing either is important it may be our range isn't used by the aliens.
