There is a word in Indian Bengali which is "sala", but in Banladesh Bengali it is pronounced as "Hala". The "s" becomes "h" in a Bangladeshi's tongue. Similarly "Tsunami" seems to be impossible to be pronounced (ts part); but it is asssumable Japanese can utter it. So why does such things happen? Is it because those who can utter hard constructs, have a letters in their alphabet which helps them pronouncing? Or is it totally a regional and anthropological thing?
A child, or more precisely an infant (from Latin in-fans - not-speaking), is said to be "universal phonetician". This stems from the fact that you can observe in children of this age the ability to distinguish between any two phonemes in any world language (and the child can grow and learn to make that difference in speech).
Now this is no supernatural ability, superpower or proof that our educational systems kill children's wonderful abilities. This means that the child is a rough, unformed piece of wood that will eventually evolve into a precisely carved... something (sorry, I had a beautiful metaphor in mind but it slipped away).
During this process of learning, you put stuff into context and you discard the rest, you build your phonological space, you segment the vowel space as required by the realities of language environment around you and you train your articulatory muscles for the gestures required to utter the sounds of that language.
After this lengthy process is accomplished, you will inevitably struggle with distinctions, which you were learning in the long previous years to be irrelevant (because in your language they are varieties of the same phoneme). So you have to unlearn what you learned and overcome the constraints set up by your own own phonology. And considering certain articulatory gestures have to be very, very precise to produce the required sound (e.g. Czech voiced /h/, voiced /ř/ etc.), it may require lot of effort to learn them past certain age.
Every language has it own phonology and rules for putting them together. Every language has something called phonotactics, which are language specific rules for what sounds can occur, where they can occur and in what combinations they can occur. For example, English phonotactics prevent the consonant cluster [pf], whereas in German [pf] is acceptable as in the word Pferd 'horse'. Also in English we pronounce the word gnome [nom] instead of [gnom], because English phonotactics prevents the cluster [gn]. It is a result of these rules that some languages can have certain sounds and combinations of sounds that other languages can't have.