The honest answer depends on the audio and on who or what is doing the typing.
Manual transcription
A widely used rule of thumb is about four hours of work for one hour of clear audio with two speakers, for an experienced typist. Beginners, poor audio, heavy accents, technical vocabulary or several overlapping speakers can push that to six hours or more.
AI transcription
AI produces a draft in a fraction of the recording length. In Transcribela, the speed depends on your computer and the quality setting. With graphics acceleration in Chrome or Edge, an hour of audio often takes only a few minutes on the Balanced and Accurate settings. Without it, the work runs on the processor and takes longer.
Proofreading the draft
For anything you will publish or quote, listen and correct. Proofreading usually takes somewhere between the audio length and twice the audio length. Clicking timestamps to jump to doubtful passages, instead of replaying everything, saves a lot of that time.
What slows things down
- Distance from the microphone and room echo.
- People talking over each other.
- Background music or noise.
- Uncommon names and jargon. Add them to your corrections list once.
Try the transcription time calculator to estimate your own workload.
Try it now: transcribe a file free, with nothing uploaded.