You don't need years to produce a phoneme that can be recognized by a native speaker. The important point is two-fold: being close enough to the actual sound, and working on minimal pairs where confusion can occur. It means you don't need to sweat over how you pronounce /r/ in French because however you make it it won't be confused with something else, but one sure should work on making their Chinese /t/ and /tʰ/ correctly. Also for listening given the sheer amount of audio and video that there is on the web it is easy to cram the required hours of listening in a few months.