We introduce SpeechEditBench, a bilingual benchmark for instruction-guided speech editing. It covers seven atomic editing tasks—content, speaker, emotion, style, prosody, paralinguistic, and acoustic editing—as well as compositional editing with multiple instructions in one sample. The benchmark uses anchor-based evaluation to separately measure target success, content preservation, and joint success, revealing where current Speech LLMs fail beyond target-only metrics.</p>\n<p>Code and data are available at:<br><a href=\"https://github.com/daxintan-cuhk/SpeechEditBench\" rel=\"nofollow\">https://github.com/daxintan-cuhk/SpeechEditBench</a><br><a href=\"https://huggingface.co/datasets/DiscreteSpeech/SpeechEditBench\">https://huggingface.co/datasets/DiscreteSpeech/SpeechEditBench</a></p>\n","updatedAt":"2026-06-05T00:23:26.506Z","author":{"_id":"6458b6508aa54fb020f5c66c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6458b6508aa54fb020f5c66c/1zlrnOOTBhZBznb1o9qLj.jpeg","fullname":"DarcyTan","name":"Darcy0123","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.800849199295044},"editors":["Darcy0123"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/6458b6508aa54fb020f5c66c/1zlrnOOTBhZBznb1o9qLj.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2606.01804","authors":[{"_id":"6a2216393490a593e87b11ca","name":"Hanlin Zhang","hidden":false},{"_id":"6a2216393490a593e87b11cb","name":"Daxin Tan","hidden":false},{"_id":"6a2216393490a593e87b11cc","name":"Dehua Tao","hidden":false},{"_id":"6a2216393490a593e87b11cd","name":"Xiao Chen","hidden":false},{"_id":"6a2216393490a593e87b11ce","name":"Haochen Tan","hidden":false},{"_id":"6a2216393490a593e87b11cf","name":"Linqi Song","hidden":false}],"publishedAt":"2026-06-03T00:00:00.000Z","submittedOnDailyAt":"2026-06-04T00:00:00.000Z","title":"SpeechEditBench: A Bilingual Multi-Attribute Benchmark for Instruction-Guided Speech Editing","submittedOnDailyBy":{"_id":"6458b6508aa54fb020f5c66c","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6458b6508aa54fb020f5c66c/1zlrnOOTBhZBznb1o9qLj.jpeg","isPro":false,"fullname":"DarcyTan","user":"Darcy0123","type":"user","name":"Darcy0123"},"summary":"Instruction-guided speech editing requires a model to modify specified speech attributes while preserving unrelated characteristics. Despite rapid progress in Speech Large Language Models (Speech LLMs), systematic evaluation of this capability remains challenging, as existing benchmarks are fragmented across isolated editing tasks. To bridge this gap, we introduce SpeechEditBench, a bilingual multi-attribute benchmark for instruction-guided speech editing. SpeechEditBench encompasses seven atomic editing tasks, as well as compositional editing tasks that integrate multiple operations within a single instruction. We propose an anchor-based evaluation protocol that separately assesses the edit success of target attributes and the preservation of untargeted attributes, leading to three metrics: target success, preservation success, and joint success. Using this benchmark, we evaluate mainstream Speech LLMs and specialized speech editing systems. The results reveal three key findings: (1) no single model performs well across all editing dimensions; (2) closed-source Speech LLMs generally outperform open-source models; (3) compositional editing remains highly challenging, with even the most advanced models struggling to achieve high joint success. SpeechEditBench provides a rigorous diagnostic framework to identify bottlenecks in Speech LLMs, thereby facilitating the development of next-generation Speech LLMs with more robust and precise instruction-guided editing capabilities. Data and code are avaialble at https://github.com/daxintan-cuhk/SpeechEditBench .","upvotes":0,"discussionId":"6a22163a3490a593e87b11d0","githubRepo":"https://github.com/daxintan-cuhk/SpeechEditBench","githubRepoAddedBy":"user","ai_summary":"A bilingual multi-attribute benchmark for instruction-guided speech editing is introduced to systematically evaluate speech modification capabilities across atomic and compositional tasks.","ai_keywords":["Speech Large Language Models","speech editing","instruction-guided editing","bilingual benchmark","atomic editing tasks","compositional editing tasks","anchor-based evaluation protocol","target success","preservation success","joint success"],"ai_summary_model":"Qwen/Qwen2.5-Coder-32B-Instruct","githubStars":1,"organization":{"_id":"5f83c275f0801648bf88454a","name":"huawei-noah","fullname":"HUAWEI Noah's Ark Lab","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1602470452594-5f83c19ff0801648bf884549.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[],"acceptLanguages":["en"],"organization":{"_id":"5f83c275f0801648bf88454a","name":"huawei-noah","fullname":"HUAWEI Noah's Ark Lab","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/1602470452594-5f83c19ff0801648bf884549.png"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2606/2606.01804.md"}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2606.01804 in a model README.md to link it from this page.

Cite arxiv.org/abs/2606.01804 in a Space README.md to link it from this page.

Add this paper to a collection to link it from this page.