Transcription

What transcription involves

Turning recorded speech into accurate text. Automatic transcription has improved enormously and is now good enough for clean, single-speaker audio in a major language. It still falls apart on the things that matter most: several people talking over each other, strong regional accents, specialist terminology, and recordings made in real rooms rather than studios.

Those are exactly the situations where an accurate transcript is worth paying for, so we transcribe with people rather than software, and use software only as a first pass that a human then corrects.

What we transcribe

• Interviews and research sessions
• Focus groups and multi-speaker discussions
• Documentary and reality footage
• Legal, medical and technical recordings
• Meetings, hearings and conference sessions
• Podcast episodes for publishing and SEO

Verbatim or clean read

There are two useful kinds of transcript and they serve different purposes. Verbatim keeps every hesitation, repetition and false start, which matters for research, legal and any use where exactly what was said is the point. Clean read removes filler and repairs sentence structure, which is what you want if the transcript is going to be read or published. Tell us which you need, because converting between them later is extra work.

Timecoding and speaker labels

For anything that will be edited, we timecode at whatever interval suits your workflow and label speakers consistently throughout. Editors working from a well-timecoded transcript find selects far faster than by scrubbing footage.

Multilingual work

We transcribe in over 60 languages with native transcribers, and can deliver a transcript in the original language, a translated transcript, or both side by side.

Send the audio with the turnaround you need and we will quote.