Text-to-speech converts written text into spoken language, making it the counterpart of speech recognition (speech-to-text). Modern systems use neural networks to produce natural-sounding voices with appropriate intonation. TTS is used in telephone announcements, navigation devices, read-aloud functions for accessible websites and voice assistants.
Speech recognition, also called ASR (automatic speech recognition), automatically converts spoken language into written text. Modern systems use neural networks trained on large amounts of audio recordings and cope reasonably well with accents and background noise. Applications include dictation, subtitles, voice assistants and the transcription of meetings and customer calls. Technical terms, proper names and poor audio quality remain common sources of error.
Audio transcription converts spoken language from recordings into written text. Modern automatic speech recognition systems also add punctuation, distinguish between speakers and sometimes provide timestamps. It is used for meeting minutes, interviews, subtitles and documenting customer calls. Quality depends heavily on audio quality, accent and specialist vocabulary; for conversations involving personal data, consent and data protection must be considered.
Natural language generation (NLG) is the area of language processing that produces understandable text from data, bullet points or other inputs. Earlier NLG systems relied on templates and rules; today most use large language models. Typical uses include weather and stock market reports, product descriptions, summaries of key figures and chatbot answers. It is important that figures and facts from the inputs are carried over correctly and unchanged.
