Glossary
Automatic speech recognition (ASR)
Automatic speech recognition (ASR) is software that converts spoken audio into written words. Modern systems are neural networks trained on large amounts of transcribed speech, and they run in the cloud or on the device.
Updated October 17, 20267 cited facts, each traced to its source
Also called ASR, speech-to-text, STT.
Automatic speech recognition (ASR) is the task of converting spoken audio into text with a computer. Product menus and API names usually call it speech-to-text (STT). Dictation software is the everyday form of ASR: you speak, and the words appear in the document. The input is audio, and the output is a sequence of words. 1
ASR turns sound into probable word sequences
An ASR system first converts the audio into a compact numerical description of its sound energy over time. It then searches for the word sequence that best fits both that description and the patterns of the language. In older systems these were separate components: an acoustic model, a pronunciation lexicon and a language model. Since about 2012, deep neural networks have replaced much of the acoustic part 2, and many current systems map audio directly to text in one network. Our guide to how speech recognition works walks through each stage.
The older design used hidden Markov models to handle the timing of speech, and Gaussian mixture models to score the sound 3. That history explains the names you will still see in the literature, such as hidden Markov model and acoustic model.
A concrete example: one spoken sentence
Suppose you say "recognize speech" into a microphone. The sounds in those two words are close to the sounds of other phrases, so the acoustic evidence alone is not decisive. The system weighs the sound against how likely each word sequence is in the language, and picks the most probable transcript. The second half of that job is done by a language model.
How accurate ASR is
ASR is now useful for many practical tasks, but transcription by any speaker in any environment is still far from solved 4. Accuracy is usually reported as a word error rate. In one 2025 test of a single podcast episode, Apple's speech API had a word error rate of 10.3 percent, against 1.5 percent for Whisper Large V3 Turbo 5 6. One episode is a small sample, so treat that result as a single data point. Accuracy depends on the audio, the speaker, the microphone and the vocabulary. Our overview of speech recognition accuracy explains how to read the figures.
Why ASR matters for dictation and typing
Dictation is ASR plus the step that places the text into a document. Speech can be much faster than typing for composing text. In a controlled study on an iPhone, English text entered by speech reached 153 words per minute, against 52 words per minute on the built-in keyboard 7. The study used short phrases in a quiet room, so a long document or a noisy office will produce different numbers.
The choice between cloud and on-device recognition changes privacy, latency and cost, but not the basic method. Our guide to on-device speech recognition compares the two.
Related terms
- Word error rate (WER): the standard accuracy measure for ASR.
- Real-time factor (RTF): how fast a system runs compared with the length of the audio.
- Voice activity detection (VAD): finding where speech starts and stops.
- Connectionist temporal classification (CTC): a training method for neural ASR that aligns audio with text.
- RNN transducer (RNN-T): a streaming neural architecture for ASR.
Guides that cover it
Evidence
Every number on this page, with where it comes from. Open a fact to see the exact words in its source.
- 1.
- 2.
- 3.
- 4.The 2026 draft of Jurafsky and Martin's textbook says that automatic speech recognition has matured to the point where it is viable for many practical tasks, while automatic transcription by any speaker in any environment is still far from solved.
Daniel Jurafsky and James H. Martin, Stanford University, online manuscript of the third edition, 2026 · Original
- 5.
- 6.
- 7.
Related terms
- Speech Recognition: What It Is and How Well It WorksThe topic this term belongs to
- Acoustic modelAn acoustic model is the part of a speech recognizer that scores how well a short window of audio matches the sounds of speech. Since about 2012 most acoustic models are deep neural networks, and their training data strongly affects accuracy.
- ConformerA Conformer is a neural encoder for speech recognition that combines self-attention, which sees the whole input, with convolution, which captures local sound patterns. FastConformer is NVIDIA's faster variant of the same design, used in streaming models.
- Connectionist temporal classification (CTC)Connectionist temporal classification (CTC) is a way to train a neural network to turn audio into text without knowing in advance which moment of sound matches which letter. The network labels every short slice of audio with a character or a blank, and the blanks and repeated labels are then removed to leave the transcript.
- Hidden Markov model (HMM)A hidden Markov model (HMM) is a statistical model in which a sequence of hidden states produces the observations you can measure, and each state changes to the next with set probabilities. In speech recognition, the hidden states are units of sound and the observations are short frames of audio.
- Inverse text normalization (ITN)Inverse text normalization (ITN) is the step that converts a speech recognizer's spoken-form output into written form, so that 'two hundred' becomes 200. It changes how the text reads, not which words the recognizer heard.
- Language model (in speech recognition)In speech recognition, a language model (LM) scores how probable a word sequence is in a language. The recognizer combines that score with the acoustic evidence to choose between words that sound alike, such as homophones.
- LibriSpeechLibriSpeech is a free corpus of about 1,000 hours of read English speech, taken from LibriVox audiobooks, and it is one of the most widely used benchmarks for English speech recognition. Its scores describe clean, read speech, which differs from everyday dictation.
- Mel spectrogramA Mel spectrogram is a grid of sound energy over time, with frequency bands spaced to match human hearing. Many speech recognizers compute its logarithm, the log-Mel spectrogram, and feed it to a neural network instead of the raw waveform.
- Real-time factor (RTF)The real-time factor (RTF) is the time a system takes to process audio divided by the length of that audio. An RTF of 1 or lower means the system keeps up with real-time speech. Some leaderboards report the inverse, RTFx, where higher is faster.
- RNN transducer (RNN-T)An RNN transducer (RNN-T) is a neural speech model that reads audio in order and emits words as it goes, so text can appear while the speaker is still talking. It has an encoder for the sound, a prediction network for the words already written, and a joint network that chooses the next output.
- Speaker diarizationSpeaker diarization is the task of working out who spoke when in a recording, by splitting the audio into turns and grouping turns that come from the same voice. It is separate from transcribing the words, and a transcript gains speaker labels only when diarization is run alongside recognition.
- Voice activity detection (VAD)Voice activity detection (VAD) is a component that decides whether a stretch of audio contains speech. Speech recognizers use it to split audio into turns, skip silence and decide when a speaker has finished.
Pages that use this term
- How Speech Recognition Works, From Sound to TextGuide
- How to Transcribe Audio to TextGuide
- What Is Dictation Software? Types and How It WorksGuide
- What Is Dictation? Voice Typing and Talk to Text ExplainedGuide
- Is Speech to Text AI? What Kind of AI Voice Typing UsesGuide
- What Is Whisper AI? OpenAI's Speech Model ExplainedGuide
- Speech-to-Text APIs Compared: Price, Accuracy, SpeedGuide
- Video to Text: Transcribe Videos and YouTubeGuide
- Word Error Rate (WER): How to Calculate and Read ItGuide
- AI Medical Scribes and Ambient Clinical DocumentationGuide
- Speech Recognition Models: How Neural ASR WorksGuide
- Transcribe Audio With Whisper, ChatGPT or GeminiGuide
- Home Assistant Voice: Speech to Text and Voice CommandsGuide
- Open-Source Dictation Software: Free Apps on GitHubGuide
- Speech Recognition in Python: Libraries ComparedGuide
- Are AI summaries of voice recordings accurate?Question
- Can DaVinci Resolve transcribe audio for free?Question
- Can Whisper AI do text to speech?Question
- Can you transcribe audio offline without uploading it?Question
- Do cloud speech-to-text APIs store your audio?Question
- Does speech recognition understand Swiss German?Question
- Does Whisper AI have an API?Question
- How do you add speech to text to an Android or iOS app?Question
- How do you get a Google Speech-to-Text API key?Question