# Speaker Recognition and Diarization

> Speaker recognition from GMM-UBM and i-vectors to learned embeddings and diarization, including threshold calibration, channel mismatch and overlap.

- Author: Muhammet Ali Köker
- Language: en
- Canonical: https://alikoker.com.tr/en/speaker-recognition-and-diarization
- Translation: https://alikoker.com.tr/seste-konusmaci-tanima
- Published: 2022-08-05T15:59:04+03:00
- Modified: 2026-08-31T22:21:00+03:00
- Verified: 2026-08-31T22:07:00+03:00
- Type: article

Speaker recognition produces identity-related similarity scores from speech samples. Identification, verification and diarization are different decision problems.

## Classical approaches

GMM-UBM systems compare a speaker distribution with a universal background model. The i-vector approach represents speech in a compact total-variability space. Methods such as PLDA score similarity between these representations.

## Learned representations

Modern systems use x-vectors and related [speaker embeddings](/en/wiki/speaker-embedding). Training encourages samples from the same speaker to be close while separating different speakers.

[Speaker recognition](/en/wiki/speaker-recognition) converts a continuous score into accept or reject decisions using a threshold. Channel mismatch, codec, utterance duration and background noise can shift score distributions.

## Speaker diarization

[Speaker diarization](/en/wiki/speaker-diarization) answers “who spoke when?” by combining VAD, segment embeddings and clustering. Unknown speaker count and overlapping speech remain difficult conditions.

A speaker score is not a deterministic identity marker; it should be interpreted together with threshold calibration, recording conditions and measured error rates.

## Cite This Work

Köker, M. A. (2022). Speaker Recognition and Diarization. alikoker.com.tr. https://alikoker.com.tr/en/speaker-recognition-and-diarization

- BibTeX: https://alikoker.com.tr/en/speaker-recognition-and-diarization.bib
- RIS: https://alikoker.com.tr/en/speaker-recognition-and-diarization.ris
- CSL-JSON: https://alikoker.com.tr/en/speaker-recognition-and-diarization.csl.json