A recording of a lecture is useful, but it is a terrible thing to revise from. You cannot search it, skim it, or highlight a definition in it. Text you can read, edit, and search is worth far more than the same hour sitting in an audio file, and turning one into the other is mostly a matter of working in short sections.
Record so the transcript has a chance
Put the phone close to the speaker, in a quiet room, and let one person speak at a time. Mumbling from the back row, background music, and three people talking over each other defeat every transcription method, human or automatic. If you are recording a voice note to yourself, speak at a normal pace and say numbers and names clearly.
Convert in short sections
- Record with Voice Recorder, or use the recording you already have.
- Open Speech to Text, pick your language, and dictate notes to text on your phone without typing, a few minutes at a time: play a section in a quiet room and let the tool listen, or re-speak the key points in your own words.
- Save the text as you go, section by section, instead of trying to do an hour in one pass.
Clean it up before you trust it
Names, numbers, dates, and technical terms are the first things any transcription gets wrong. Read through once, fix those, add headings, and pull out the definitions and deadlines. The transcript is a draft, not a finished set of notes, and the ten minutes spent fixing it are what make it usable.
Speech to Text listens live through your browser, so it works best in a supported browser with microphone permission, and it does not take an audio file and hand back a transcript in one click. Working section by section is slower, but it is the workflow that actually finishes.
The result is notes you can search the night before the exam, instead of an hour of audio you will never replay. That is the whole point of converting: not a perfect transcript, but usable notes.