Is speech to text the same as dictation?

Not exactly. Speech to text is the technology: software that converts spoken words into written text. Dictation is one use of that technology, the one where you speak your own words and they appear where you are writing. So all dictation is speech to text, but not all speech to text is dictation. The what is dictation guide gives the full definition, and this page explains where the two terms split apart.

Dictation is one use of speech to text

Speech to text is the wider name. Microsoft's Azure Speech service groups speech to text with text to speech and other capabilities as parts of one product ¹. Microsoft also lists a call-transcription use case for the same service: transcribe calls in real time or process a batch of calls ². A call transcript is speech to text, but it is not dictation. Nobody is writing their own words, and nothing is being typed into a document as it is spoken.

Dictation adds three conditions. The speaker is the author. The words go straight into the message, note or document being written. And the author reads and corrects the text as it appears. Remove any one of these and you have a different job, which is why the dictation and transcription question separates them.

The product names mix the two terms

Companies name the same feature differently, and that is the main source of confusion. Apple calls it Dictation. Microsoft calls it voice typing on Windows. Samsung ships two voice-to-text apps on its phones, Samsung voice input and Google Voice Typing ³. On a Mac, Apple's guide says Dictation lets you speak to enter text anywhere you can type it ⁴.

The engine underneath is also called speech to text. Google's help page for Google Docs says the web browser controls the speech-to-text service behind its voice typing ⁵. Windows voice typing uses online speech recognition powered by Azure Speech services ⁶.

Use the table as a guide to the names, not as a rule. A product can call its feature whatever it likes.

Three cases where speech to text is not dictation

The clearest differences show up in three situations:

1. Live captions or meeting notes. The software writes what other people say. You are not the author, so this is not dictation.
2. Transcribing a recording. The audio already exists, and the output is a record of what was said. This is transcription, not dictation.
3. Voice commands. Saying "delete that" or "open Mail" tells the computer to act. Dictation understands a few commands for writing, but running the computer by voice is voice control, a different job.

It depends: the same tool can do both jobs

Some tools blur the line, so check what you are using:

• Recorded dictation. You can speak notes into a phone's voice recorder and get a transcript later. You spoke the words, but the transcript is made from a recording, so the output is transcription. A university help article says that on an iPhone you tap the speech bubble icon in the bottom left of a Voice Memos recording to generate the transcription ⁷.
• Dictation in an app that also transcribes files. Some apps dictate live and also transcribe audio files. The live mode is dictation, and the file mode is transcription.
• Medical dictation. Clinicians often dictate into a recorder, and the words are typed up later. The speaking step is dictation, and the typing step is transcription. The transcription overview covers the second half.

What to do: decide what the output is

Ask one question before you choose a tool: is the text something you are writing, or a record of someone else's words?

1. If you are writing it, use dictation in the app where the text goes. The how to turn on voice typing guide explains the setting on each device.
2. If you have a recording, use transcription. The transcribe audio to text guide covers the options, and the free audio to text converter works in your browser.
3. Check the output either way. Speech to text makes errors that someone has to catch. In clinical notes dictated in 2016, the speech recognition draft had a 7.4 percent error rate, or 7.4 errors per 100 words ⁸. Check the draft before you rely on it. The proofreading guide explains how to find the mistakes.

For the related terms, read the automatic speech recognition and dictation entries. The voice typing entry covers the Microsoft and Google names. If you need the other direction, where the computer reads text aloud, see speech to text vs text to speech. For the whole subject, start at the dictation overview.