An iterative phase-recovery algorithm that reconstructs a time-domain signal from a magnitude spectrogram. It alternates between inverse-STFT to estimate phase, forward-STFT to enforce the magnitude constraint, and repeats until the error between the estimated and target magnitude spectrogram falls below a convergence threshold. The algorithm is deterministic given the same initialization and parameters. Parameters: magnitude spectrogram input, maximum iteration count, convergence tolerance, and initial phase guess. Persistence mechanism: implemented in signal-processing libraries (librosa, MATLAB, Python scipy), documented in peer-reviewed literature, and used in audio synthesis and speech processing pipelines. [formal: Griffin-Lim | substrate: behavior | horizon: hours | explicit: yes | epoch: 0.01]
Accepted ontology entry
griffin-lim
An iterative phase-recovery algorithm that reconstructs a time-domain signal from a magnitude spectrogram. It alternates between inverse-STFT to estimate phase, forward-STFT to enforce the magnitude constraint, and repeats until the error…
Definition
Why it is in scope
An iterative algorithm for reconstructing a signal from its magnitude spectrogram by alternating between time-domain synthesis and frequency-domain magnitude enforcement until convergence. Human-made mathematical procedure, encoded in software and signal-processing literature.
Names and aliases
- griffin-limen · CANONICAL
Relations from this entry
- cmsr5uo3j00bzkp53ssqaygttINSTANCE_OF →
Griffin-Lim IS a specific kind of phase-retrieval: it is a concrete algorithm for recovering missing phase information from a magnitude spectrogram. A competent speaker would call Griffin-Lim a phase-retrieval algorithm. Nearest kind: phase-recovery and iterative-algorithm do not exist, so phase-retrieval is the nearest kind.
- cmsqg4mko005f3e32eyp15c4sDEPENDS_ON →
Griffin-Lim operates by alternating between inverse-STFT and forward-STFT. Remove STFT and the algorithm stops operating entirely — it cannot function without the STFT/ISTFT pair. DEPENDS_ON direction correct.
- cmsrf31nw019okp53l8esrbaeINSTANCE_OF →
Griffin-Lim is a specific phase-recovery algorithm: it is a kind of phase-recovery, not merely dependent on it. Law 9: X is a specific case of the general Y. A competent speaker would call Griffin-Lim 'a phase-recovery algorithm.' Nearest kind is phase-recovery (not a more general category).
- cmsps9i0v06eejlssqrjcqyviSERVES →
Griffin-Lim is an iterative phase-recovery algorithm designed to reconstruct audio signals from magnitude spectra. Its entire purpose is to serve signal processing — specifically audio reconstruction and speech processing. SERVES: servant (algorithm) → master (domain it serves).
- cmsr5uo3j00bzkp53ssqaygttDERIVED_FROM →
phase-retrieval as a general mathematical problem (reconstructing signals from magnitude-only measurements) predates and generalizes the Griffin-Lim algorithm. Griffin-Lim is a specific iterative algorithm developed within the phase-retrieval domain. Which-came-first test: phase-retrieval concept existed first, Griffin-Lim algorithm was developed as a method to solve it.
- cmspltmgd05tcjlsse8vqfce1DEPENDS_ON →
Removal test: Griffin-Lim algorithm iteratively computes STFT and inverse-STFT to estimate phase from magnitude. Remove STFT from the algorithm and it has no forward/backward transform mechanism — the entire iterative process collapses. STFT is not just historical context; it is the operational core.
- cmsrf31nw019okp53l8esrbaeSERVES →
Griffin-Lim is designed and maintained for the sake of phase-recovery: it reconstructs phase information from magnitude-only spectrograms, which is the core problem in phase recovery. The algorithm's entire purpose is to solve the phase-recovery problem.
- audio-processingSERVES →
Griffin-Lim is an iterative phase retrieval algorithm built and maintained for the sake of audio processing: it reconstructs time-domain audio from magnitude spectrograms, enabling audio inpainting and vocoder applications. Servant points at master.
- speech-enhancementSERVES →
Griffin-Lim iterative phase reconstruction is built for the sake of recovering perceptually plausible audio waveforms from magnitude spectrograms, a core step in speech enhancement pipelines that need phase-aware reconstruction after mask-based magnitude enhancement. Servant points at master per Law 8d.
- cmsrfbgwp01aykp53wp0b02wyDERIVED_FROM →
Griffin-Lim is a specific iterative algorithm for phase retrieval that grew out of phase vocoder principles. Phase vocoder existed first and fed into Griffin-Lim development. Which-existed-first test satisfied.
- cmsr5uo3j00bzkp53ssqaygttSERVES →
Griffin-Lim iterative algorithm is designed and maintained specifically to perform phase retrieval from magnitude spectrograms. Its purpose is phase retrieval.
Relations to this entry
No accepted relations in this direction.
Record identity
- Created
- Aug 13, 2026, 10:42 AM UTC
- Content hash
- a84724529786b2a20672eb0bc33a4ebb2b6a758929fc46122f76b2b2cbcfd513