SRT / ASS / VTT

SRT is the plain one and its simplicity is why it won: a counter, a start and end time, and the words. Any tool can write one, any player can read one, and you can fix a typo in a text editor. What it cannot do is say where on the screen a line should sit or what colour it should be.

ASS was built for the cases where that matters — a sign on a wall translated in place, two people speaking at once, karaoke where each syllable lights up. It carries fonts, colours, outlines, positions and timing effects, and an MKV can even carry the font files needed to render it as intended. A player that renders ASS as plain text is showing you the words and none of the work.

PGS and VobSub are a different thing entirely. They are pictures of subtitles, taken from Blu-ray and DVD, where the subtitle was always drawn as an image. They can be displayed and nothing more: no font change, no size change, no searching, no translating. Turning one into text requires optical character recognition and an afternoon.

On a desktop, film.mkv and film.srt in the same folder is all a player needs — the convention is that old and that universal. iOS has no such convention, and more to the point has no such folder: apps are sandboxed, each with its own storage, and "next to" is not a relationship the system exposes.

So a subtitle file on an iPhone has to be handed to a player deliberately, and the player has to be the thing that keeps the pairing. Once it does, the choice can stick to the file — a film you left with Turkish subtitles opens with them next time.

The other half of the problem is encoding. A subtitle file is bytes, and nothing in the file says which alphabet those bytes mean unless it was written as UTF-8. Older files, particularly Turkish and Cyrillic ones, were written in regional encodings, and a player that assumes UTF-8 renders ş and ğ as garbage. Detecting the encoding rather than assuming it is what separates a player that shows those files correctly from one that does not.