|
Whisper is an automatic speech recognition (ASR) system by OpenAI for transcribing speech into text; see Radford et al. (2022) for the model architecture, training and evaluation. OpenAI’s Whisper implementation is in Python (PyTorch).
Praat uses whisper.cpp, a lightweight C/C++ port of Whisper built on top of the ggml tensor library for machine learning, developed by Georgi Gerganov and many other contributors (see Acknowledgments). The original OpenAI Whisper models must be converted to ggml format for use with whisper.cpp.
For how transcription is used in Praat, see the Speech recognition tutorial. For the transcription settings, see transcription with whisper.cpp.
© Anastasia Shchupak 2026-06-01