Bengali subtitle & caption generator

Most Bengali letters hang from the matra, a headline like the one in Devanagari, but some carry only part of it and a few none at all, so the top of a word is not always an unbroken line. The vowel signs ি, ে and ৈ are written before the consonant they follow in speech, and ো and ৌ come in two pieces placed either side of it, as in কো.

Clusters stack further. Many Bengali conjuncts set one consonant on top of another, and some, জ্ঞ among them, are drawn as a shape of their own. য following a consonant turns into the ya-phala, a mark attached to its right side as in ব্যস্ত, and ু, ূ and ৃ sit below the letter.

Most Indian scripts in AutoCap fall back to one of the Baloo typefaces when a template's font cannot draw them. Bengali falls back to Noto Sans Bengali instead, loaded in bold weights. Baloo Da 2, the Baloo design for Bengali, is still in the font list for anyone who wants that rounded look. The ten Bengali faces are:

The vowel carried by a Bengali consonant is said nearer to o than to a, but transliteration often writes it as a. That is why রবীন্দ্রনাথ is spelt Rabindranath in English while it sounds closer to Robindronath, and why কথা turns up as both katha and kotha. Before correcting a romanised transcript, see which convention it follows and decide which your viewers expect.

Bengali script sidesteps that question for viewers who read it, and choosing it narrows the Family list to Bengali faces, with a link at the bottom to show the rest.

An English translation has to add something Bengali leaves unsaid: gender. সে and তিনি mean he or she alike, so the translation must pick one, and where the clip gives no clue it can pick wrongly. Read the pronouns in a translated transcript before exporting.

Bengali has its own digits, and one is easy to misread: ৪ means four but looks like a Latin 8. When a caption carries a price, a date or a phone number, compare it with the audio, and consider Latin digits if your viewers are used to them.

A Bengali sentence ends with the dari (।), so if full stops turn up in a Bengali-script transcript, replace them. Some verbs have two spellings in common use, হল and হলো among them, and the transcript is the place to make them match.