Pith. sign in

REVIEW 4 major objections 4 minor 48 references

AudioScape-TTA: A Structured Soundscape Benchmark for Fine-Grained Text-to-Audio Evaluation

T0 review · 4 major / 4 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read AudioScape-TTA argues that checking each requested sound event, attribute, and spoken phrase yields evaluations of text-to-audio models that track human judgment far better than global similarity scores do.

desk verdict Serious benchmark paper with honest controls; the rubrics-vs-CLAP result is plausible but rests on a small human study and an evaluator whose accuracy on generated audio is unmeasured. read the letter →

arxiv 2608.04479 v1 pith:E4QN3KYC submitted 2026-08-05 cs.SD cs.AI

Jinting Wang , Yuguang Yang , Shengyu Li , Yan Rong , Shan Yang , Xiaoda Yang , Li Liu This is my paper · ORCID
classification cs.SDcs.AI
keywords text-to-audiogenerationsoundscapebenchmarkrubric-basedevaluationaudio-groundedspeechcontentpreservationcomplexity-awaresemanticinstructionfollowing
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

AudioScape-TTA is a benchmark and evaluation pipeline built to test whether text-to-audio models actually follow fine-grained instructions, not just produce plausible audio. The paper assembles 2,258 real-world soundscape clips with structured annotations that separate scene, sound effects, background music, and speech, and converts each annotation into binary yes/no rubrics (25,707 total) that an audio-language model and an ASR system check against generated audio. Across 13 open-source TTA models, the rubric scores expose a persistent gap: models can generate the right events but fail to control their acoustic attributes, and nearly all models fail to preserve spoken content. The paper's key evidence is that the aggregate rubric satisfaction rate correlates with human semantic ratings at Spearman rho = 0.879, while CLAP similarity correlates at only 0.312, arguing that requirement-level evaluation is more human-aligned than global embedding similarity.

What carries the argument

The load-bearing object is the structured soundscape schema plus its derived rubric set: each clip is annotated as a scene with sound effects, background music, and speech, and the annotation hierarchy Sample→Modality→Event→Attribute is deterministically converted into binary yes/no questions that are fixed across all models. Two verifiers carry the evaluation: an audio-language model answers event-presence and event-attribute questions from the audio, and an ASR system supplies transcripts that are compared to the annotated target utterance by mixed-unit (Latin words plus Chinese characters) clipped-count coverage at a 0.60 threshold. Sample complexity is quantified along two axes, event density and structural complexity, which are weakly correlated (r = 0.221) and jointly partition the benchmark into Easy, Medium, and Hard subsets; this design is what lets the paper separate 'can generate the sound' from 'can control the sound' and from 'can compose a dense soundscape.'

What would settle it

Re-run the 13-model evaluation on the 130 user-study samples using a different audio-language verifier (or a human-annotated calibration set over generated audio) and recompute model-level Spearman correlation between Overall SR and the composite human semantic score; if the correlation drops sharply below 0.879—or if the evaluator's agreement on a held-out human-labeled set falls well below the reported 84.88%—the claim that rubric scores align with human semantic judgment is not supported. A second check: verify that the balanced sensitivity result (94.91% balanced accuracy on 452 reference items) holds when positive/negative pairs are drawn from generated audio rather than reference clips.

Watch

Extended reading notes

Core claim

The paper claims that a structured, rubric-based, audio-grounded evaluation can measure fine-grained text-to-audio instruction following more faithfully than global similarity metrics, and it supports this with a benchmark where every sample has fixed semantic criteria. The central numerical finding is that across the 13 evaluated models, Overall SR—the micro-average satisfaction of event-presence, event-attribute, and speech-content rubrics—correlates at rho = 0.879 (Holm-adjusted p = 0.0015) with a composite human semantic score, whereas CLAP_MS correlates at rho = 0.312 (p = 0.0739). The same benchmark shows that attribute satisfaction lags behind event presence for essentially all models, that 11 of 13 models achieve 0% speech-content satisfaction at the 0.60 coverage threshold, and that satisfaction rate degrades monotonically as soundscape complexity rises. The paper concludes that global embedding and perceptual-quality metrics are complementary but incomplete, and that structured rubrics provide interpretable diagnoses that better track human semantic judgment.

Load-bearing premise

The whole comparison rests on the automatic verifier providing accurate enough yes/no answers for generated audio—if the audio-language model's errors are correlated with model properties like audio quality or speech-likeness, the model rankings and the 0.879 correlation could shift; the reference-audio sanity check shows the evaluator agrees with annotation-derived answers only 84.88% of the time.

Editorial extensions

If this is right

  • Fine-grained attribute control is the current bottleneck: across 13 models, event-attribute satisfaction systematically trails event-presence satisfaction, so future TTA systems should be optimized for attribute fidelity rather than event presence alone.
  • Speech-content preservation is a separate capability from speech-like audio generation; the benchmark's near-zero SCCA@0.60 for most models shows that treating speech as a generic sound event is insufficient.
  • Model rankings under CLAP-based similarity do not reproduce rubric-based rankings; for example, MMAudio has the highest CLAP_MS (0.5169) but only 64.73% overall SR, so optimizing global alignment alone would mislead development.
  • Robustness under complexity is measurable and matters: most models degrade from Easy to Hard subsets, while the top two models hold up best in dense, speech-involved soundscapes.
  • Because every rubric is fixed and shared across models, results are auditable and can be traced to specific failures, which is a concrete advantage over a single similarity score.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural next test is to swap the verifier: replacing the paper's audio-language model with a stronger one would likely raise absolute satisfaction rates (the paper's own sanity check shows a stronger verifier reaches 95.03% agreement on reference audio), and the comparison would show whether model-level rankings are robust or verifier-dependent.
  • The same rubric machinery could be extended to temporal, spatial, and causal relations, which the paper explicitly leaves out; that would turn the benchmark from a semantic-composition test into a scene-organization test.
  • The 452-item balanced sensitivity analysis is performed on reference audio only, so the reported 94.91% accuracy should not be read as the error rate on generated audio; a human-annotated calibration set over generated clips would be the missing check on the headline correlation.
  • The 0.60 coverage threshold is lenient in an order-agnostic way, so speech-content scores are best interpreted as 'did the model produce most of the required words,' not as verbatim transcription accuracy; the paper's threshold sweep shows the ranking holds at 0.80.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper introduces AudioScape-TTA, a benchmark of 2,258 audio-text pairs with 25,707 binary question-answer rubrics, organized around a structured soundscape representation (scene, sound effects, background music, speech) and a two-dimensional complexity characterization (event density and structural complexity). The authors propose a rubric-based, audio-grounded evaluation framework in which event-presence and event-attribute rubrics are verified by Qwen3-Omni-Instruct and speech-content rubrics by Qwen3-ASR-based transcript coverage (SCCA@0.60). They evaluate 13 open-source TTA models and report that rubric-based satisfaction rates degrade with increasing complexity and reveal persistent weaknesses in attribute control and speech-content preservation. The central claim is that the rubric-based evaluation aligns better with human semantic judgments than global similarity: in a user study with 10 prompts and 13 models, the Overall SR correlates with a composite human semantic score at Spearman rho = 0.879 (Holm-adjusted p = 0.0015), whereas CLAP_MS correlates at rho = 0.312 (p = 0.0739).

Significance. If the central claim holds, this is a valuable contribution to text-to-audio evaluation. The benchmark is carefully constructed with human refinement of annotations, deterministic rubric generation, and a clear separation of the rubric-generation loop from the external human validation. The paper provides extensive supplementary analyses, including audio-masked and shuffled-audio controls, a balanced sensitivity analysis with contrastive negatives, threshold sensitivity for speech-content coverage, and answer-order randomization checks. These checks are commendable and go beyond what is typical for TTA benchmarks, making the proposed evaluation protocol potentially reusable and reproducible. The complexity-aware analysis offers a useful template for diagnosing model robustness. However, the strength of the human-alignment claim is currently limited by the small human study (10 prompts, 130 samples), the lack of verifier reliability evidence on generated audio, and the opaque statistical reporting in the headline correlation table.

major comments (4)
  1. [Table 4] The reported Holm-adjusted p-values do not match the reported Spearman coefficients and sample size. With N=13 models, Overall SR rho=0.879 corresponds to a two-sided p-value of about 0.00008 under the standard t-approximation, not 0.0015; similarly, CLAP_MS rho=0.312 corresponds to a raw p-value of about 0.30, not 0.0739. The authors must report the exact inference procedure (e.g., permutation test with a stated number of permutations, tie handling, or the precise formula) and both raw and adjusted p-values, because the headline comparison's statistical significance is not reproducible as written.
  2. [Sanity Checks (Tables 14-16)] All verifier reliability checks are performed on reference audio. The audio-masked control shows an 18.31% 'yes' rate on silence, and the balanced sensitivity analysis uses Gemini-verified reference-audio negatives; neither provides false-positive or false-negative rates on generated audio, where artifacts, missing events, and degraded speech are common. If verifier errors correlate with model properties such as audio quality or speech-likeness, the model-level correlation in Table 4 could be inflated by evaluator bias. The authors should provide a verifier-robustness analysis on generated audio (e.g., compare Qwen3-Omni-Instruct against Gemini-2.5-Pro on generated samples, or obtain human labels on a subset) and show that model rankings are stable under verifier choice.
  3. [User Study (Section 5)] The human-alignment evidence rests on 10 prompts, 130 audio samples, and 13 model-level points. Because every model is scored on the same 10 prompts, the model-level correlation could be dominated by a few prompt-specific effects. The paper does not report per-prompt correlations, bootstrap confidence intervals for rho, or any inter-rater reliability statistic (e.g., ICC or Krippendorff's alpha). To support the claim that rubric-based evaluation is more human-aligned than global similarity, the authors should either extend the human study to a larger prompt set or provide a robustness analysis showing that the correlation is stable across prompt subsets and rater subsets.
  4. [Section 5 (Human Alignment of Evaluation Metrics)] The abstract and text claim superiority over 'conventional global similarity metrics,' but Table 4 only compares against CLAP_MS. The paper computes several other global metrics (FAD, KL, ISC, Audiobox-Aesthetic CE/CU/PC/PQ) and reports them in the supplementary material, yet never correlates them with human semantic judgments. The authors should either include these correlations or narrow the claim to 'CLAP_MS' to avoid over-generalization.
minor comments (4)
  1. [Section 5 (Human Alignment of Evaluation Metrics)] The phrase 'the our rubric-based metrics' contains a typo; it should be 'our rubric-based metrics'.
  2. [Table 10 caption] The caption reads 'bestand second-bestresults' with missing spaces; please correct to 'best and second-best results'.
  3. [Supplementary WER/CER diagnostics] The subsets J_WER and J_CER are referenced as |J_WER|=69 and |J_CER|=11, but their construction is not defined before these references; please clarify how the valid reference clips are selected, especially why three of the 83 target-utterance clips are excluded.
  4. [User Study (Section 5)] The paper states that 10 prompts are 'randomly selected covering diverse modality compositions and complexity levels' but does not report the distribution of the selected prompts across the Easy/Medium/Hard partitions or modality compositions. Given the small sample, reporting this distribution is important for assessing representativeness.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the headline human-alignment result is validated against independent human ratings, and no fitted parameter is relabeled as a prediction.

full rationale

AudioScape-TTA's derivation chain is self-contained rather than circular. The rubric QA items are generated from the human-refined structured annotations, and the reference answer is set to 'A' by construction because each question is a positive requirement derived from an annotated event or attribute; however, this is a benchmark-construction step that fixes oracle labels on reference audio only, not a predicted quantity. The paper's central validation (Table 4) uses an independent external signal: 26 participants' ratings of 130 generated samples on event completeness and attribute satisfaction, aggregated into a composite human semantic score. No parameter is fitted to those human ratings; the speech threshold tau=0.60 is fixed in advance, its sensitivity is examined in Table 11 across thresholds 0.40-0.80, and answer-position robustness is checked in Tables 16-17. The self-references that appear in the related-work discussion (TTA-Bench, AudioAtlas, T2A-Feedback/EpicBench, AnyAudio-Judge) are comparative citations, not load-bearing justifications, so even if some authors overlap with the present paper, no central premise depends on an unverified prior claim. The appendix caveat that Gemini's higher agreement does not 'definitively separate evaluator errors from rubric-construction errors' is an honest validity limitation about verifier accuracy on reference audio, not a reduction of the central result to its inputs. Consequently, the reported human-alignment advantage of rubrics over CLAP is an external empirical comparison rather than a self-fulfilling definition.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The benchmark relies on hand-designed complexity measures, a hand-chosen speech threshold, and strong assumptions about annotation quality and automatic verifier reliability. No new physical or computational entities are postulated; the complexity metrics are definitions rather than inventions. The most load-bearing assumptions are the accuracy of the human annotations and the automatic evaluator, which the paper only partially validates.

free parameters (4)
  • SCCA coverage threshold tau = 0.60
    Hand-selected threshold converting order-agnostic mixed-unit coverage into binary speech-content satisfaction. The sensitivity analysis in Table 11 shows ranking stability across tau=0.40-0.80.
  • Complexity partition thresholds = Easy: C<=-1; Medium: -1<C<1; Hard: C>=1
    Hand-chosen z-score cutoffs in Eq. (4) to keep sufficient samples per subset. Not fitted to model performance, but affects all complexity-level conclusions.
  • Event density weighting = Density = N_SFX + N_Speech + N_BGM
    Hand-designed counting convention in Eq. (1) counting SFX instances but treating speech and BGM as binary presence indicators. A modeling choice that affects the Easy/Medium/Hard split.
  • Structural complexity weighting = Structure_Score = S_modality + S_cooccur (range 0-4)
    Hand-designed measure in Eqs. (2)-(3) that assigns one point per active modality plus a speech co-occurrence bonus. This encodes an assumption that speech co-occurrence adds complexity.
assumptions (5)
  • domain assumption Human annotations on the 2,258 clips are semantically correct and consistent; trained annotators are the final authority.
    Stated in Data Collection and Annotation, but no inter-annotator agreement or annotation error rate is reported. All benchmark rubrics inherit this assumption.
  • domain assumption Qwen3-Omni-Instruct and Qwen3-ASR provide sufficiently accurate audio-grounded verification of event presence, attributes, and speech content.
    Used in Audio-grounded Verification. The reference-audio sanity check shows 84.88% Overall SR, implying about 15% verifier disagreement with annotation-derived answers; the gap varies with evaluator choice (Gemini reaches 95.03%).
  • domain assumption Order-agnostic clipped-count transcript coverage (R_mixed) at threshold 0.60 captures whether intended spoken content is preserved.
    Defined in Eqs. (8)-(10). It does not penalize word-order changes or extra transcript units, so it measures unit coverage rather than exact content preservation.
  • domain assumption Movie/TV extracted clips are a valid proxy for realistic soundscapes.
    The benchmark is constructed from movie and television productions; the domain distribution may not generalize to all text-to-audio use cases.
  • standard math z-score normalization and indicator-function aggregation are standard statistical operations.
    Used in Eqs. (4)-(5) without further justification, which is reasonable for a benchmark construction.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AudioScape-TTA: A Structured Soundscape Benchmark for Fine-Grained Text-to-Audio Evaluation." pith.science (2026). https://pith.science/paper/E4QN3KYC

@misc{pith2026260804479,
  author       = {Pith},
  title        = {Pith review of: AudioScape-TTA: A Structured Soundscape Benchmark for Fine-Grained Text-to-Audio Evaluation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/E4QN3KYC}},
  note         = {Machine review of arXiv:2608.04479}
}
read the original abstract

Text-to-audio (TTA) generation has recently achieved remarkable progress in synthesizing realistic audio from natural language descriptions. However, determining whether generated audio faithfully satisfies complex textual instructions remains challenging. Existing benchmarks mainly rely on global similarity metrics, providing limited insight into fine-grained semantic failures. To address this limitation, we introduce \textbf{AudioScape-TTA}, a structured and complexity-aware benchmark for fine-grained TTA evaluation. AudioScape-TTA represents realistic soundscapes through modality-aware semantic structures and characterizes generation complexity using event density and structural complexity. Based on these annotations, we propose a rubric-based audio-grounded evaluation framework that verifies event realization, acoustic attributes, and speech content through fine-grained semantic criteria. The benchmark contains 2,258 audio-text pairs with 25,707 binary QA rubrics, enabling scalable and interpretable analysis of TTA systems. Experiments on 13 representative open-source TTA models reveal persistent limitations in fine-grained attribute control, speech-content preservation, and compositional soundscape generation. Human validation further demonstrates that our rubric-based evaluation achieves stronger alignment with human semantic judgments than conventional global similarity metrics.

Figures

Figures reproduced from arXiv: 2608.04479 by the authors.

Figure 1
Figure 1. Comprehensive evaluation of open-source TTA generation models on the AudioScape-TTA benchmark. (a) Semantic [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of AudioScape-TTA, a complexity-aware benchmark for text-driven omni-audio generation. The top panel illustrates the benchmark construction and evaluation pipeline. The bottom panel characterizes AudioScape-TTA in a complexity-aware manner by organizing samples within the proposed Density–Structure complexity space. For each candidate clip, we adopt a semi-automatic an￾notation pipeline in which an LLM is u… view at source ↗
Figure 3
Figure 3. User study results on the AudioScape-TTA bench [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Statistics of the proposed AudioScape-TTA benchmark. (a) Density distribution reflects the combination of discrete SFX-instance counts and clip-level presence indicators for speech and BGM. (b) Structure score distribution characterizes the complexity and organization …

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 20 canonical work pages

  1. [1]

    Communication, Simulation, and Intelligent Agents: Implications of Personal Intelligent Machines for Medical Education

    Clancey, William J. Communication, Simulation, and Intelligent Agents: Implications of Personal Intelligent Machines for Medical Education. Proceedings of the Eighth International Joint Conference on Artificial Intelligence (IJCAI-83)

  2. [2]

    Classification Problem Solving

    Clancey, William J. Classification Problem Solving. Proceedings of the Fourth National Conference on Artificial Intelligence

  3. [3]

    , title =

    Robinson, Arthur L. , title =. 1980 , doi =. https://science.sciencemag.org/content/208/4447/1019.full.pdf , journal =

    1980

  4. [4]

    New Ways to Make Microcircuits Smaller---Duplicate Entry

    Robinson, Arthur L. New Ways to Make Microcircuits Smaller---Duplicate Entry. Science

  5. [5]

    Clancey and Glenn Rennels , abstract =

    Diane Warner Hasling and William J. Clancey and Glenn Rennels , abstract =. Strategic explanations for a diagnostic consultation system , journal =. 1984 , issn =. doi:https://doi.org/10.1016/S0020-7373(84)80003-6 , url =

    doi:10.1016/s0020-7373(84)80003-6 1984

  6. [6]

    and Rennels, Glenn R

    Hasling, Diane Warner and Clancey, William J. and Rennels, Glenn R. and Test, Thomas. Strategic Explanations in Consultation---Duplicate. The International Journal of Man-Machine Studies

  7. [7]

    Poligon: A System for Parallel Problem Solving

    Rice, James. Poligon: A System for Parallel Problem Solving

  8. [8]

    Transfer of Rule-Based Expertise through a Tutorial Dialogue

    Clancey, William J. Transfer of Rule-Based Expertise through a Tutorial Dialogue

Show all 48 references
  1. [9]

    The Engineering of Qualitative Models

    Clancey, William J. The Engineering of Qualitative Models

  2. [10]

    2023 , eprint=

    Attention Is All You Need , author=. 2023 , eprint=

    2023

  3. [11]

    Pluto: The 'Other' Red Planet

    NASA. Pluto: The 'Other' Red Planet

  4. [12]

    arXiv preprint arXiv:2209.15352 , year=

    Audiogen: Textually guided audio generation , author=. arXiv preprint arXiv:2209.15352 , year=

    arXiv

  5. [13]

    International Conference on Machine Learning , pages=

    AudioLDM: Text-to-Audio Generation with Latent Diffusion Models , author=. International Conference on Machine Learning , pages=. 2023 , organization=

    2023

  6. [14]

    International Conference on Machine Learning , pages=

    Make-an-audio: Text-to-audio generation with prompt-enhanced diffusion models , author=. International Conference on Machine Learning , pages=. 2023 , organization=

    2023

  7. [15]

    arXiv preprint arXiv:2305.18474 , year=

    Make-an-audio 2: Temporal-enhanced text-to-audio generation , author=. arXiv preprint arXiv:2305.18474 , year=

    arXiv

  8. [16]

    IEEE/ACM Transactions on Audio, Speech, and Language Processing , volume=

    Auffusion: Leveraging the power of diffusion and large language models for text-to-audio generation , author=. IEEE/ACM Transactions on Audio, Speech, and Language Processing , volume=. 2024 , publisher=

    2024

  9. [17]

    EzAudio: Enhancing Text-to-Audio Generation with Efficient Diffusion Transformer , author=. Proc. Interspeech 2025 , pages=

    2025

  10. [18]

    Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    Meanaudio: Fast and faithful text-to-audio generation with mean flows , author=. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

  11. [19]

    arXiv preprint arXiv:2412.21037 , year=

    Tangoflux: Super fast and faithful text to audio generation with flow matching and clap-ranked preference optimization , author=. arXiv preprint arXiv:2412.21037 , year=

    arXiv

  12. [20]

    ACM Multimedia 2024 , year=

    Tango 2: Aligning diffusion-based text-to-audio generative models through direct preference optimization , author=. ACM Multimedia 2024 , year=

    2024

  13. [21]

    arXiv preprint arXiv:2606.03672 , year=

    Foley-Omni: A Unified Multimodal Generation Model from Task-Level Audio Synthesis to Complete Video Soundtrack Generation , author=. arXiv preprint arXiv:2606.03672 , year=

    arXiv

  14. [22]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Omni2sound: Towards unified video-text-to-audio generation , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  15. [23]

    Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

    Audiostory: Generating long-form narrative audio with large language models , author=. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages=

  16. [24]

    Audiocaps: Generating captions for audios in the wild , author=. Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) , pages=

    2019

  17. [25]

    ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Clotho: An audio captioning dataset , author=. ICASSP 2020-2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2020 , organization=

    2020

  18. [26]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Tta-bench: A comprehensive benchmark for evaluating text-to-audio models , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  19. [27]

    Proceedings of the 33rd ACM International Conference on Multimedia , pages=

    AudioAtlas: A Comprehensive and Balanced Benchmark Towards Movie-Oriented Text-to-Audio Generation , author=. Proceedings of the 33rd ACM International Conference on Multimedia , pages=

  20. [28]

    Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing , pages=

    RiTTA: Modeling Event Relations in Text-to-Audio Generation , author=. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing , pages=

    2025

  21. [29]

    ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Audiotime: A temporally-aligned audio-text benchmark dataset , author=. ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2025 , organization=

    2025

  22. [30]

    arXiv preprint arXiv:2509.17765 , year=

    Qwen3-omni technical report , author=. arXiv preprint arXiv:2509.17765 , year=

    arXiv

  23. [31]

    arXiv preprint arXiv:2601.21337 , year=

    Qwen3-ASR Technical Report , author=. arXiv preprint arXiv:2601.21337 , year=

    arXiv

  24. [32]

    International Conference on Learning Representations , volume=

    Masked audio generation using a single non-autoregressive transformer , author=. International Conference on Learning Representations , volume=

  25. [33]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Mmaudio: Taming multimodal joint training for high-quality video-to-audio synthesis , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  26. [34]

    ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Stable audio open , author=. ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2025 , organization=

    2025

  27. [35]

    arXiv preprint arXiv:2605.27838 , year=

    Dasheng AudioGen: A Unified Model for Generating Coherent Audio Scenes from Text , author=. arXiv preprint arXiv:2605.27838 , year=

    arXiv

  28. [36]

    2025 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) , pages=

    Meta audiobox aesthetics: Unified automatic assessment for speech, music and sound , author=. 2025 IEEE Automatic Speech Recognition and Understanding Workshop (ASRU) , pages=. 2025 , organization=

    2025

  29. [37]

    ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Natural language supervision for general-purpose audio representations , author=. ICASSP 2024-2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2024 , organization=

    2024

  30. [38]

    ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=

    Large-scale contrastive language-audio pretraining with feature fusion and keyword-to-caption augmentation , author=. ICASSP 2023-2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , pages=. 2023 , organization=

    2023

  31. [39]

    IEEE/ACM Transactions on Audio, Speech, and Language Processing , volume=

    Diffsound: Discrete diffusion model for text-to-sound generation , author=. IEEE/ACM Transactions on Audio, Speech, and Language Processing , volume=. 2023 , publisher=

    2023

  32. [40]

    arXiv preprint arXiv:2312.15821 , year=

    Audiobox: Unified audio generation with natural language prompts , author=. arXiv preprint arXiv:2312.15821 , year=

    arXiv

  33. [41]

    arXiv preprint arXiv:2310.00704 , year=

    Uniaudio: An audio foundation model toward universal audio generation , author=. arXiv preprint arXiv:2310.00704 , year=

    arXiv

  34. [42]

    Proceedings of the 31st ACM international conference on multimedia , pages=

    Text-to-audio generation using instruction guided latent diffusion model , author=. Proceedings of the 31st ACM international conference on multimedia , pages=

  35. [43]

    Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    T2A-Feedback: Improving Basic Capabilities of Text-to-Audio Generation via Fine-grained AI Feedback , author=. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

  36. [44]

    Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

    Vintage: Joint video and text conditioning for holistic audio generation , author=. Proceedings of the Computer Vision and Pattern Recognition Conference , pages=

  37. [45]

    arXiv preprint arXiv:2412.15922 , year=

    RiTTA: Modeling event relations in text-to-audio generation , author=. arXiv preprint arXiv:2412.15922 , year=

    arXiv

  38. [46]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Audio generation with multiple conditional diffusion model , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

  39. [47]

    arXiv preprint arXiv:2606.03116 , year=

    AnyAudio-Judge: A Dynamic Rubric-Based Benchmark and Evaluator for Audio Instruction Following , author=. arXiv preprint arXiv:2606.03116 , year=

    arXiv

  40. [48]

    arXiv preprint arXiv:2601.14728 , year=

    AQAScore: Evaluating Semantic Alignment in Text-to-Audio Generation via Audio Question Answering , author=. arXiv preprint arXiv:2601.14728 , year=

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.