Urdu to Roman Transliteration

In short Transliterating Urdu into Roman letters is the easier direction, because Urdu script is far more regular than Roman Urdu: each letter has a defined sound and the mapping is largely one to one. Diacritics are stripped first, aspirated consonants are handled as digraphs — kh, gh, th, dh and the rest, which are single letters in Urdu written with the do-chashmi he — and the result is readable. Two limits are worth stating because they are properties of the script rather than of the tool. The retroflex letters ٹ ڈ ڑ have no standard Roman marking, so they are written as t, d and r like their dental counterparts, and the distinction is lost. And short vowels cannot be recovered at all: Urdu writes them as diacritics that are omitted in ordinary text, so کتب gives no clue whether it is kutub or katab. No transliterator can recover information the text never carried, which is why the output needs a reader who knows the words. Transliteration is not translation — the sounds are converted, the meaning is not.