Turkmen subtitling services

By reading the line against the audio and by not cutting the words that anchor it. A Turkmen subtitler hears the long vowel even though the page does not show it, so the check is whether a viewer with only the text in front of them can still land on the right word. Where a cue cannot carry that much, we rebuild the cue rather than shorten the sentence.

Usually not. Turkmen readers do not expect them and ordinary published text does not use them, so macrons in a subtitle read as a textbook rather than as dialogue. There are exceptions worth having: language teaching, pronunciation material, anything aimed at learners. If that is your project, we mark length the same way throughout and say so on the delivery note.

Because the model is choosing between two words that look the same and only differ in how long a vowel is held. Even when it hears the difference, the spelling it outputs is identical, so an error is invisible in the file and only shows up when a Turkmen reader meets a sentence about grass that should have been about fire. A person fixes those, in context.

Because a glossary matches strings, and two Turkmen words that differ only by vowel length are the same string. A terminology check will happily confirm a term that is not the term, and a search across a series will return both. We deal with it the way a reader does, by looking at what surrounds the word, which is why terminology on a Turkmen project is checked by a linguist rather than a report.

Twenty pounds a minute is where it starts. Running time accounts for most of the total, then transcription if the audio has never been typed up, an accessibility version or a burnt-in master if you want one, and the number of delivery formats. The context work described on this page is inside the rate, not an extra line. Every quote comes itemised.

Yes, and that is where this page's point does the most work. Captions in the same language put the viewer in front of Turkmen text with no second language to cross-check against, so a line that leaves two readings open stays open. We write those lines so the context survives, and the speaker labels and audio cues go in as they would in any SDH job.

Send it. Our subtitlers time it to picture and rework the line breaks, and the pass that matters comes after: reading the cues cold, as a viewer would, and marking any that could be taken two ways. That is quick on a file written carefully and slow on one that was machine-generated, and either way it beats starting over. Where nothing exists, we transcribe the audio here.

They do, in SRT, VTT, STL or TTML, into Premiere, Resolve, Final Cut, Avid and the caption tools the platforms ship with. The Turkmen letters that carry marks sit in the Latin ranges that ordinary subtitle fonts already cover, so what we ask you to look at is the preview rather than the import log. A burnt-in master is there if you want the question closed.