AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Blow: a single-scale hyperconditioned flow for non-parallel raw-audio voice conversion

Joan Serrà, Santiago Pascual, Carlos Segura Perales

NeurIPS 2019 · Probabilistic and scientific modeling › Molecules and proteins · approach: Flow matching generative modeling · 49 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use Blow instead of VQ-VAENon-parallel voice conversion on VCTK · VCTK testSpoofing [%]: 65.0 -> 89.3Improved
Use Blow instead of StarGANNon-parallel voice conversion on VCTK · VCTK testSimilarity [%]: 61.8 -> 77.6Improved
Use Blow instead of StarGANNon-parallel voice conversion on VCTK · VCTK testNaturalness [1--5]: 2.87 -> 2.83No clear effect
Use Blow instead of Glow-WaveNetFlow-based voice conversion on VCTK · VCTK testL [nat/dim]: 4.18 -> 4.45Improved
Use Blow instead of Glow-WaveNetFlow-based voice conversion on VCTK · VCTK testSpoofing [%]: 3.1 -> 89.3Improved
Replace Blow's 8×12 structure with 3×32 (ablation)Blow · VCTK, 5 min per speaker, 100 epochsSpoofing [%]: 66.2 -> 17.2Hurt
Use 3×32 structure with squeeze of 8 (ablation)Blow · VCTK, 5 min per speaker, 100 epochsSpoofing [%]: 66.2 -> 65.7No clear effect
Use a multi-scale structure (ablation)Blow · VCTK, 5 min per speaker, 100 epochsSpoofing [%]: 66.2 -> 3.5Hurt
Use multi-scale structure with 5×19 flows and squeeze of 4 (ablation)Blow · VCTK, 5 min per speaker, 100 epochsSpoofing [%]: 66.2 -> 16.6Hurt
Replace hyperconditioning with additive conditioning in the coupling network (ablation)Blow · VCTK, 5 min per speaker, 100 epochsSpoofing [%]: 66.2 -> 39.5Hurt
Replace hyperconditioning with additive conditioning before ActNorm (ablation)Blow · VCTK, 5 min per speaker, 100 epochsSpoofing [%]: 66.2 -> 22.5Hurt
Remove data augmentation (ablation)Blow · VCTK, 5 min per speaker, 100 epochsSpoofing [%]: 66.2 -> 28.3Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy