Romanization refers to the process of conversion of writing from a non-Roman script into the Roman (Latin) script, or a method of implementing it. The process of Romanizing a text is divided into transliteration for written documents, transcription for spoken texts, and combinations of both. There are several consistent or standardized romanization systems. They can be classified based on their characteristics. The characteristics of a system may allow it to be used for various, sometimes contradictory applications like document retrieval, linguistic analysis, easy readability, and accurate representation of pronunciation.
Transliteration attempts to align the source script with the target script one-to-one, with less emphasis on how the result sounds when pronounced in the language of the reader. For instance, using the Sata srī akāla romanization of Punjabi, an informed reader is able to reconstruct the original Punjabi script with 100% accuracy, but having a good grasp of pronunciation requires additional knowledge. A transliteration is the process of mapping from one system of writing to another based on phonetic similarity. This tool converts Latin letters (such as a, b, c, etc.) into characters with a similar pronunciation in the target language. For example, in Hindi transliteration, you can type in “namaste” to get “नमस्ते”, which sounds like “namaste”. “Transliteration” is different from “Translation”. In Transliteration the conversion is based on the pronunciation, and not just the meaning.
The purpose of most Romanizations is to make the source language accessible to casual readers who are unfamiliar with the original script. These romanizations are based on phonemic transcription, which attempts to accurately convey the significant sounds (phonemes) in the target language. Using phonetic conversion, all phonetic sounds in the source language can be represented, possibly at the expense of legibility, when using characters or conventions which are not found in the target language.