Yoruba transcription services

Both are correct Yoruba, so we set the convention with you rather than picking by default. Fused spelling — ra ẹja written as rẹja — matches how the phrase actually sounds and reads naturally to a Yoruba speaker. Writing the two words in full is clearer for anyone learning the language, for subtitling where every word needs to map to timing, or for a legal record where the fuller form removes ambiguity. We apply whichever you choose consistently through the whole transcript.

Because the sound genuinely no longer contains two separate words. When ra ẹja becomes rẹja in speech, the vowel boundary a speech-to-text model would need to detect the split has physically merged into one sound. The model has to guess where a word ends without acoustic evidence, and Yoruba's tone marking on top of that gives it another axis to get wrong at the same time.

Possibly, and that is expected rather than a quality problem. Standard Yoruba spelling generally follows the contracted, elided form of speech, but an optional apostrophe convention exists for marking where the elision happened, and not everyone uses it. What we guarantee is not one universally "correct" spelling of every elided pair, but a convention chosen for your project and held consistently from the first line to the last.

It is mostly a matter of what ends up on the page rather than what the sentence means — unlike, say, Yoruba's tone marks, which do change meaning. The risk with elision is legibility and consistency: a transcript that fuses some word pairs and spells others out in full reads as inconsistent rather than wrong, and a subtitle file where fused words disrupt reading speed or line breaks needs a human decision about when to unpack them.

There is no flat rate, because no two files take the same work. Runtime, recording quality, how many people are talking, your deadline and whether you also need an English version all move the figure. Every quote carries our Price Match Promise.

Upload is encrypted, the file goes to one named Yoruba transcriber who has signed an NDA, and nothing is put through public speech-to-text services at any point. If you have your own confidentiality agreement we sign it, if you have retention rules we work to them, and if your policy says the audio goes once the transcript is approved, we delete it.

It is normal work. Phone audio, site noise, focus groups and meetings where three people speak at once all get transcribed. What difficult audio changes is time, and occasionally certainty about a word or a name. We listen to a sample first and tell you straight what we expect to be able to capture, rather than handing you a file peppered with inaudible tags.

In whatever the next step needs: plain text, Word, PDF, SRT, VTT or your own template, with timecodes and speaker labels set the way you want them. Tonal marks and the fused or unfused spelling of elided words are delivered in Unicode so they display correctly wherever the file lands. Tell us what happens to the transcript after we send it and we format it for that, not for us.