DEFINE: Exemplar-Guided Accent Control for Zero-Shot TTS

Hugging Face Daily Papers Papers

Summary

DEFINE is an end-to-end zero-shot TTS framework that decouples speaker identity from accent by conditioning each on separate audio exemplars, with an inference-time guidance weight controlling accent strength on top of F5-TTS using LoRA adaptation.

Zero-shot text-to-speech (TTS) can reproduce an unseen speaker from a short reference recording, but typically entangles speaker identity and accent within the same reference. We introduce DEFINE, an end-to-end framework that decouples these factors by conditioning speaker identity and target accent on separate audio exemplars. A single inference-time guidance weight continuously controls accent strength without retraining. Built on F5-TTS with parameter-efficient LoRA adaptation, DEFINE maps short accent exemplars into a conditioning space using an exemplar encoder supervised through learned accent prototypes, requiring neither accent labels at inference time nor post-synthesis waveform conversion. On seen accents, increasing accent guidance improves accent-probe accuracy from 6.5% to 19.6%. More importantly, a single DEFINE model generalizes accent control beyond its training accent set: on seen and out-of-domain accents, though not on held-out accents, it matches the accent transfer performance of a two-model TTS-voice-conversion cascade while achieving higher speaker similarity and comparable predicted speech quality. These results demonstrate that speaker identity and accent can be independently controlled from audio exemplars within a single zero-shot TTS model, including for accents unseen during training.
Original Article
View Cached Full Text

Cached at: 10/05/26, 04:46 AM

Paper page - DEFINE: Exemplar-Guided Accent Control for Zero-Shot TTS

Source: https://huggingface.co/papers/2609.32777

Abstract

Zero-shottext-to-speech(TTS)canreproduceanunseenspeakerfromashortreferencerecording,buttypicallyentanglesspeakeridentityandaccentwithinthesamereference.WeintroduceDEFINE,anend-to-endframeworkthatdecouplesthesefactorsbyconditioningspeakeridentityandtargetaccentonseparateaudioexemplars.Asingleinference-timeguidanceweightcontinuouslycontrolsaccentstrengthwithoutretraining.BuiltonF5-TTSwithparameter-efficientLoRAadaptation,DEFINEmapsshortaccentexemplarsintoaconditioningspaceusinganexemplarencodersupervisedthroughlearnedaccentprototypes,requiringneitheraccentlabelsatinferencetimenorpost-synthesiswaveformconversion.Onseenaccents,increasingaccentguidanceimprovesaccent-probeaccuracyfrom6.5%to19.6%.Moreimportantly,asingleDEFINEmodelgeneralizesaccentcontrolbeyonditstrainingaccentset:onseenandout-of-domainaccents,thoughnotonheld-outaccents,itmatchestheaccenttransferperformanceofatwo-modelTTS-voice-conversioncascadewhileachievinghigherspeakersimilarityandcomparablepredictedspeechquality.Theseresultsdemonstratethatspeakeridentityandaccentcanbeindependentlycontrolledfromaudioexemplarswithinasinglezero-shotTTSmodel,includingforaccentsunseenduringtraining.

View arXiv pageView PDFProject pageGitHub2Add to collection

Get this paper in your agent:

hf papers read 2609\.32777

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.32777 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.32777 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.32777 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Edge0/Audio8-TTS-Realistic-Air

Hugging Face Models Trending

Edge0/Audio8-TTS-Realistic-Air is an Apache-2.0 text-to-speech model finetuned from Edge0/Audio8-TTS-Preview-0.6b, supporting zero-shot voice cloning and multilingual synthesis across 25+ languages (zh, en, ja, ko, de, fr, es, and more).