Metraipolis › archive › dc_ea03f103216a

Rethinking Automated Voice Similarity by Shifting from EER to Embedding Geometry — arXiv:2609.33999

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.33999
License (as archived)
CC0-1.0
Archived
2026-10-08T01:37:17.249Z
SHA-256 of the archived text
93e26c598c48f09c0ca4b0926fd26e009db8350278e5bb6426eb2fc0f646c775
Stable URL of this page
https://metraipolis.com/documento/dc_ea03f103216a

How to cite

Rethinking Automated Voice Similarity by Shifting from EER to Embedding Geometry — arXiv:2609.33999 — arXiv (en), https://arxiv.org/abs/2609.33999 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ea03f103216a (Metraipolis archive dc_ea03f103216a, sha256 93e26c598c48f09c0ca4b0926fd26e009db8350278e5bb6426eb2fc0f646c775)
BibTeX
@misc{dcea03f103216a,
  title = {Rethinking Automated Voice Similarity by Shifting from EER to Embedding Geometry — arXiv:2609.33999},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33999},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ea03f103216a; sha256 93e26c598c48f09c0ca4b0926fd26e009db8350278e5bb6426eb2fc0f646c775},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33999. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Speaker verification (SV) models are commonly assumed to better capture nuances among speaker characteristics as verification accuracy improves, leading to their widespread use as automated proxies for human voice similarity in speech generation tasks. However, by establishing a human perceptual alignment metric and conducting systematic analysis, we demonstrate that perceptual alignment is governed far more by how a model is trained (its learning objective) than by how well it performs (EER). Notably, standard margin-based classification losses (e.g., AAM-Softmax) yield substantially lower perceptual alignment than prototypical metric losses, while EER itself fails to track human judgment, directly challenging the community's implicit assumption. We trace this divergence to embedding geometry, where a model's effective dimensionality ($d_{\mathrm{eff}}$) tracks perceptual alignment with a $-0.95$ rank correlation, revealing that the dimensional spread favored by classification losses fundamentally clashes with the low-dimensional nature of human voice perception. Imposing a dimensionality bottleneck compresses $d_{\mathrm{eff}}$ and raises perceptual alignment ($\rho_{\mathrm{align}}$) from 0.08 to 0.74, establishing a principled geometric criterion for evaluating voice similarity.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in eess.AS).
Categorie arXiv: Audio and Speech Processing (eess.AS) ; Sound (cs.SD).
Pagina del lavoro: https://arxiv.org/abs/2609.33999 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T01:11:22.064Z dalla pagina https://arxiv.org/catchup/eess.AS/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).