Pashto subtitle generator

Pashto subtitles that survive the dialect.

Generate an editable Pashto caption track, then review the handful of words where a Peshawar speaker and a Kandahar speaker sound nothing alike, and export a subtitle file or a captioned video.

One spelling, several pronunciations

Pashto has a quirk that makes automatic captioning harder than it looks, and it is not the alphabet. Two of its letters are pronounced in genuinely different ways depending on where the speaker comes from.

One of them is a retroflex sh sound in Kandahar and closer to kh in Peshawar. The other is a retroflex zh in Kandahar and closer to a hard g in Peshawar. Quetta and Ghazni sit at other points again. Speakers from those places are saying the same word and writing it the same way, but a listener trained on one region hears something else entirely.

Automatic transcription is that listener. A model whose Pashto came mostly from one dialect will confidently produce a different word for a speaker from another, and because the output is a real Pashto word rather than nonsense, nothing about it looks wrong on the page.

What this means in practice

The good news

Written Pashto stays largely consistent across dialects. So there is no decision to make about which form to publish, and no need for separate tracks per region. The correct caption is the same either way.

The work

Because the failure is in hearing rather than spelling, the fix is a targeted review rather than a rewrite. Check the words carrying those letters, and leave the rest of the draft alone.

A review order that suits the problem

Start with the dialect-sensitive words, because that is where a mis-hearing becomes a wrong word rather than a typo. Then names of people, places and tribes, which are what viewers write in to correct. Then Urdu, Dari or English words spoken inside Pashto sentences, which usually belong exactly as they were said rather than respelled into Pashto.

Line length comes last, and it matters more than it does in English, because Perso-Arabic Pashto sets long in a vertical frame. If a line runs past the safe area, the subtitle line breaker will re-break it without touching the timing, and the subtitle validator will tell you which cues are over the limit in the first place.

Fonts, files, and where the track goes

Pashto uses the Perso-Arabic family but adds letters Urdu and Arabic do not have, including a set of retroflex consonants. A font built only for Urdu may have no glyph for them, so a few characters render as boxes while the rest of the line looks fine. Confirm the file is UTF-8 first, using the subtitle encoding guide, then try a second player before concluding the text is wrong.

YouTube takes a subtitle track per language, so a Pashto track can sit on the same upload as others; the subtitle track upload guide covers that path. Short-form feeds take no subtitle file at all, so those need captions rendered into the picture, as the delivery guide explains.

For neighbouring languages, see the Urdu subtitle generator, the Sindhi subtitle generator, and the Punjabi caption generator.

Pashto subtitle questions

Why does Pashto transcription struggle more than Urdu?

Because the same written letter is pronounced differently depending on where the speaker is from. The letter that sounds like a retroflex sh in Kandahar sounds closer to kh in Peshawar, and the letter that sounds like a retroflex zh in Kandahar is closer to a hard g in Peshawar. A model that learned one of those pronunciations mis-hears the other, even though both speakers would write the word identically.

Does the spelling change between dialects?

Generally no, and that is the useful part. Written Pashto stays largely consistent while the pronunciation varies, so the correct output is usually the same regardless of the speaker's dialect. The failure is in recognition, not in spelling, which means the fix is a review pass rather than a decision about which form to write.

What should I check first in a Pashto caption draft?

Anywhere a dialect-sensitive letter appears, since that is where a mishearing becomes a different word rather than a typo. Proper nouns next, especially tribal and place names. Then anything spoken in Urdu or English mid-sentence, which needs one consistent treatment across the file.

Does Pashto use the same alphabet as Urdu?

Related, but not interchangeable. Pashto carries several letters that neither Urdu nor Arabic uses, so software built for those languages can be missing them outright. If a keyboard or font cannot produce a character, no amount of careful transcription will put it in the file.

Why do only some characters turn into boxes?

That is a font problem rather than a broken file. Confirm the file is saved as UTF-8 first, then open it in a different player. If the text is correct there, the original player's font simply has no glyph for the Pashto-specific letters.

What can I export?

Reviewed Pashto captions export as SRT or VTT subtitle files, plain text, or a captioned video. Keep the subtitle file even when you ship a captioned video, because it is the version you can still correct later. Open the caption generator.

Start a Pashto caption track

Transcribe the video, check the dialect-sensitive words and the names, then export SRT, VTT, text, or a captioned video.