ARXIV:2605.30748 · TEXT-TO-SPEECH · SUBMITTED 01 JUN · 20:20 UTC · FRESHNESS STALE

VerifiedSource: PDF linkedVerifiedPaperPack: citation fields availablePartialProof: unverified proof status

Chatterbox-Flash: Prior-Calibrated Block Diffusion for Streaming Zero-Shot TTS

Deokjin Seo · Gangin Park · Kihyun Nam · arXiv

A streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency.

Ship in 2-4 weeks›Score8.0Evidence unverified

Opportunity summary

Pain A streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency.

Evidence 0 refs | 4 sources | 67% coverage

Blocker Evidence unverified

Open Build Read PDF Signal Canvas Track

PROBLEM

A streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency. We find that naively transferring mainstream block-diffusion decoding to discrete speech tokens degrades quality, as a long-tail token distribution…

METHOD

Full abstract

We present Chatterbox-Flash, a zero-shot text-to-speech model obtained by fine-tuning a pretrained autoregressive TTS decoder into a block-diffusion decoder, enabling parallel token generation within each block while retaining block-by-block streaming. We find that naively transferring mainstream block-diffusion decoding to discrete speech tokens degrades quality, as a long-tail token distribution biases parallel position selection toward a few high-frequency tokens. To mitigate this without architectural modification, we introduce two inference-time techniques: prior-calibrated scoring, which subtracts the block-level marginal token distribution, and an early-decoding schedule, which adaptively terminates iteration based on calibrated confidence. On standard zero-shot TTS benchmarks, Chatterbox-Flash attains high-fidelity synthesis comparable to strong autoregressive and non-autoregressive baselines, while supporting streaming inference with time-to-first-packet on par with streaming AR systems and substantially lower real-time factor. Code and audio samples are available at https://github.com/resemble-ai/chatterbox-flash.

RESULT

ScienceToStartup currently rates this 8.0/10 on the public viability pass. Code and audio samples are available at https://github.com/resemble-ai/chatterbox-flash. A public repository is linked, so build verification can inspect implementation evidence instead of treating the…

WHY NOW

Text-to-Speech moved forward this cycle; last verified June 2026. Public score 8.0/10. Implementation evidence is present through a linked repository.

Continue into Read for claims, analysis, references, and neighboring papers.

Opportunity summary

Score8.0

PainA streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency.

Evidence0 refs | 4 sources | 67% coverage

Blockerno shell-level blocker reported

Analysis summary

A streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency.

VerifiedSource: PDF linkedVerifiedPaperPack: citation fields availablePartialProof: unverified proof status

Competitive landscape

A streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency.

Segment

Text-to-Speech

Adoption evidence

Public code linked for build inspection

Commercial read

8.0/10 public viability

Direct

not classified

Adjacent

not classified

Substitute

not classified

Unknown

not classified

{ "contract_version": "paper-r2", "paper_id": "c6cc8ee7-9d27-4892-b848-e2bfd742aa2c", "arxiv_id": "2605.30748", "canonical_route": "/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts", "active_tab": "synced from current hash by the drawer client", "selected_artifact": "chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts", "endpoints": { "paper_pack": "/api/v1/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts/paper-pack", "build_passport": "/api/v1/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts/build-passport", "mcp_resource": "sciencetostartup://surfaces/paper-workspace" } }

{ "surface": "paper", "mode": "paper", "query": "Chatterbox-Flash: Prior-Calibrated Block Diffusion for Streaming Zero-Shot TTS", "normalized_query": "2605.30748", "route": "/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts", "paper_ref": "chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts", "topic_slug": null, "benchmark_ref": null, "dataset_ref": null }

{ "@context": "https://schema.org", "@graph": [ { "@type": "WebPage", "@id": "https://sciencetostartup.com/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts#webpage", "url": "https://sciencetostartup.com/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts", "name": "Chatterbox-Flash: Prior-Calibrated Block Diffusion for Streaming Zero-Shot TTS", "description": "A streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency.", "isPartOf": { "@id": "https://sciencetostartup.com/#website" } }, { "@type": "ScholarlyArticle", "@id": "https://sciencetostartup.com/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts#scholarlyArticle", "headline": "Chatterbox-Flash: Prior-Calibrated Block Diffusion for Streaming Zero-Shot TTS", "description": "A streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency.", "url": "https://sciencetostartup.com/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts", "sameAs": "https://arxiv.org/abs/2605.30748", "identifier": { "@type": "PropertyValue", "propertyID": "arXiv", "value": "2605.30748" }, "isAccessibleForFree": true, "isPartOf": { "@id": "https://sciencetostartup.com/#website" }, "datePublished": "2026-05-29T02:25:02.000Z", "author": [ { "@type": "Person", "name": "Deokjin Seo" }, { "@type": "Person", "name": "Gangin Park" }, { "@type": "Person", "name": "Kihyun Nam" } ], "codeRepository": "https://github.com/resemble-ai/chatterbox-flash", "additionalProperty": [ { "@type": "PropertyValue", "propertyID": "viabilityScore", "value": 8 }, { "@type": "PropertyValue", "propertyID": "researchDomain", "value": "Text-to-Speech" }, { "@type": "PropertyValue", "propertyID": "commercialReadiness", "value": "code, repo url" } ] }, { "@type": "SoftwareSourceCode", "@id": "https://sciencetostartup.com/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts#software", "name": "Chatterbox-Flash: Prior-Calibrated Block Diffusion for Streaming Zero-Shot TTS - Source Code", "description": "A streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency.", "codeRepository": "https://github.com/resemble-ai/chatterbox-flash", "url": "https://github.com/resemble-ai/chatterbox-flash" }, { "@type": "BreadcrumbList", "itemListElement": [ { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://sciencetostartup.com" }, { "@type": "ListItem", "position": 2, "name": "Text-to-Speech", "item": "https://sciencetostartup.com/topics" }, { "@type": "ListItem", "position": 3, "name": "Chatterbox-Flash: Prior-Calibrated Block Diffusion for Strea", "item": "https://sciencetostartup.com/paper/chatterbox-flash-prior-calibrated-block-diffusion-for-streaming-zero-shot-tts" } ] } ] }

Competitive landscape

A streaming zero-shot text-to-speech model uses prior-calibrated block diffusion for high-fidelity synthesis with low latency.

Segment

Text-to-Speech

Adoption evidence

Public code linked for build inspection

Commercial read

8.0/10 public viability

Direct

not classified

Adjacent

not classified

Substitute

not classified

Unknown

not classified

Chatterbox-Flash: Prior-Calibrated Block Diffusion for Streaming Zero-Shot TTS

Chatterbox-Flash: Prior-Calibrated Block Diffusion for Streaming Zero-Shot TTS

Claim map

Constellation map

Competitive landscape

Buzz

PDF

REFERENCES

Related Papers

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline

Claim map

Constellation map

Competitive landscape

Buzz

PDF

REFERENCES

Related Papers

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline