mlsanityData to 7 Oct 2026

Tango 2: Aligning Diffusion-based Text-to-Audio Generations through Direct Preference Optimization

arXiv 2404.09956 ↗First posted 15 Apr 2024cs.SD cs.AI cs.CLcited by 97

On the map: Diffusion models → Audio, speech and music

also in 1 more
Abstract (arXiv)

Generative multimodal content is increasingly prevalent in much of the content creation arena, as it has the potential to allow artists and media personnel to create pre-production mockups by quickly bringing their ideas to life. The generation of audio from text prompts is an important aspect of such processes in the music and film industry. Many of the recent diffusion-based text-to-audio models focus on training increasingly sophisticated diffusion models on a large set of datasets of prompt-audio pairs. These models do not explicitly focus on the presence of concepts or events and their temporal ordering in the output audio with respect to the input prompt. Our hypothesis is focusing on how these aspects of audio generation could improve audio generation performance in the presence of limited data. As such, in this work, using an existing text-to-audio model Tango, we synthetically create a preference dataset where each prompt has a winner audio output and some loser audio outputs for the diffusion model to learn from. The loser outputs, in theory, have some concepts from the prompt missing or in an incorrect order. We fine-tune the publicly available Tango text-to-audio model using diffusion-DPO (direct preference optimization) loss on our preference dataset and show that it leads to improved audio output over Tango and AudioLDM2, in terms of both automatic- and manual-evaluation metrics.

Key facts. Cited by 97 papers in the map (data through 7 Oct 2026). Placed in Audio, speech and music, one of 14 directions in Diffusion models. Builds on 18 of its 36 references that are in the map. First posted on arXiv 15 Apr 2024.

Builds on

18 of 36 references.

Closest ancestors · 4

Same direction, newest first

  1. 2023Mustango: Toward Controllable Text-to-Music Generation14 Nov
  2. 2023AudioLDM 2: Learning Holistic Audio Generation with Self-supervised Pretraining10 Aug
  3. 2023Text-to-Audio Generation using Instruction-Tuned LLM and Latent Diffusion Model24 Apr
  4. 2023Make-An-Audio: Text-To-Audio Generation with Prompt-Enhanced Diffusion Models30 Jan

Other references · 14

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  2. 2020Denoising Diffusion Implicit Modelscited by 9,655
  3. 2016Image-to-Image Translation with Conditional Adversarial Networkscited by 4,510
  4. 2020HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesiscited by 1,139
  5. 2023Diffusion Model Alignment Using Direct Preference Optimizationcited by 767
  6. 2021SoundStream: An End-to-End Neural Audio Codeccited by 756
  7. 2022Large-scale Contrastive Language-Audio Pretraining with Feature Fusion and Keyword-to-Caption Augmentationcited by 707
  8. 2023Towards General Text Embeddings with Multi-stage Contrastive Learningcited by 705
  9. 2022AudioLM: a Language Modeling Approach to Audio Generationcited by 611
  10. 2021W2v-BERT: Combining Contrastive Learning and Masked Language Modeling for Self-Supervised Speech Pre-Trainingcited by 340
  11. 2023Efficient Diffusion Training via Min-SNR Weighting Strategycited by 242
  12. 2017Learning from Between-class Examples for Deep Sound Recognitioncited by 53
2 more
  1. 2021Decoupling Magnitude and Phase Estimation with Deep ResUNet for Music Source Separationcited by 48
  2. 2024BATON: Aligning Text-to-Audio Model with Human Preference Feedbackcited by 18

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 29; 2025: 36; 2026: 32.

97 papers.

Sort:

By area:

  1. 2024MMAudio: Taming Multimodal Joint Training for High-Quality Video-to-Audio Synthesiscited by 134
  2. 2025Ovi: Twin Backbone Cross-Modal Fusion for Audio-Video Generationcited by 76
  3. 2024TangoFlux: Super Fast and Faithful Text to Audio Generation with Flow Matching and Clap-Ranked Preference Optimizationcited by 67
  4. 2024Video-Guided Foley Sound Generation with Multimodal Controlscited by 49
  5. 2026MOVA: Towards Scalable and Synchronized Video-Audio Generationcited by 39
  6. 2024Online DPO: Online Direct Preference Optimization with Fast-Slow Chasingcited by 38
  7. 2024Align Anything: Training All-Modality Models to Follow Instructions with Language Feedbackcited by 38
  8. 2025Kling-Foley: Multimodal Diffusion Transformer for High-Quality Video-to-Audio Generationcited by 38
  9. 2025A Survey of Direct Preference Optimizationcited by 35
  10. 2024Taming Data and Transformers for Audio Generationcited by 28
  11. 2024A Comprehensive Survey of Direct Preference Optimization: Datasets, Theories, Variants, and Applicationscited by 27
  12. 2024Both Ears Wide Open: Towards Language-Driven Spatial Audio Generationcited by 25
38 more
  1. 2025AudioX: A Unified Framework for Anything-to-Audio Generationcited by 25
  2. 2025KAD: No More FAD! An Effective and Efficient Evaluation Metric for Audio Generationcited by 24
  3. 2025LeVo: High-Quality Song Generation with Multi-Preference Alignmentcited by 23
  4. 2024Align-SLM: Textless Spoken Language Models with Reinforcement Learning from AI Feedbackcited by 22
  5. 2024FlowSep: Language-Queried Sound Separation with Rectified Flow Matchingcited by 21
  6. 2024Preference Alignment Improves Language Model-Based TTScited by 21
  7. 2024AV-Link: Temporally-Aligned Diffusion Features for Cross-Modal Audio-Video Generationcited by 21
  8. 2024Improving Text-To-Audio Models with Synthetic Captionscited by 19
  9. 2024FlashAudio: Rectified Flows for Fast and High-Fidelity Text-to-Audio Generationcited by 19
  10. 2024A Survey of Sustainability in Large Language Models: Applications, Economics, and Challengescited by 18
  11. 2024Images that Sound: Composing Images and Sounds on a Single Canvascited by 17
  12. 2024ETTA: Elucidating the Design Space of Text-to-Audio Modelscited by 17
  13. 2024VinTAGe: Joint Video and Text Conditioning for Holistic Audio Generationcited by 16
  14. 2024SoPo: Text-to-Motion Generation Using Semi-Online Preference Optimizationcited by 14
  15. 2024AudioEditor: A Training-Free Diffusion-Based Audio Editing Frameworkcited by 13
  16. 2025Zero-Shot Audio-Visual Editing via Cross-Modal Delta Denoisingcited by 13
  17. 2025AudioGen-Omni: A Unified Multimodal Diffusion Transformer for Video-Synchronized Audio, Speech, and Song Generationcited by 11
  18. 2025Omni-Reward: Towards Generalist Omni-Modal Reward Modeling with Free-Form Preferencescited by 11
  19. 2026Audio-Omni: Extending Multi-modal Understanding to Versatile Audio Generation and Editingcited by 10
  20. 2025JAM: A Tiny Flow-based Song Generator with Fine-grained Controllability and Aesthetic Alignmentcited by 9
  21. 2026Omni2Sound: Towards Unified Video-Text-to-Audio Generationcited by 8
  22. 2024Synthio: Augmenting Small-Scale Audio Classification Datasets with Synthetic Datacited by 7
  23. 2025AniME: Adaptive Multi-Agent Planning for Long Animation Generationcited by 7
  24. 2024MixEval-X: Any-to-Any Evaluations from Real-World Data Mixturescited by 6
  25. 2024Challenge on Sound Scene Synthesis: Evaluating Text-to-Audio Generationcited by 6
  26. 2025Inversion-DPO: Precise and Efficient Post-Training for Diffusion Modelscited by 6
  27. 2025Latent Swap Joint Diffusion for 2D Long-Form Latent Generationcited by 5
  28. 2025MMMG: a Comprehensive and Reliable Benchmark for Multitask Multimodal Generationcited by 5
  29. 2025Generative Audio Language Modeling with Continuous-valued Tokens and Masked Next-Token Predictioncited by 5
  30. 2025Vevo2: A Unified and Controllable Framework for Speech and Singing Voice Generationcited by 5
  31. 2025AudioMoG: Guiding Audio Generation with Mixture-of-Guidancecited by 5
  32. 2025UniFlow-Audio: Unified Flow Matching for Audio Generation from Omni-Modalitiescited by 5
  33. 2025MR-FlowDPO: Multi-Reward Direct Preference Optimization for Flow-Matching Text-to-Music Generationcited by 5
  34. 2026Stable Audio 3cited by 5
  35. 2025DRAGON: Distributional Rewards Optimize Diffusion Generative Modelscited by 4
  36. 2025Speechless: Speech Instruction Training Without Speech for Low Resource Languagescited by 4
  37. 2025Direct Preference Optimization for Speech Autoregressive Diffusion Modelscited by 4
  38. 2026LeVo 2: Stable and Melodious Song Generation via Hierarchical Representation Modeling and Progressive Post-Trainingcited by 4

50 most cited of 97 All 97 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026Diffusion Models for Joint Audio-Video Generation
  2. 2025Audio Generation Through Score-Based Generative Modeling: Design Principles and Implementation
  3. 2025Diffused Responsibility: Analyzing the Energy Consumption of Generative Text-to-Audio Diffusion Models
  4. 2023ERNIE-Music: Text-to-Waveform Music Generation with Diffusion Models
  5. 2025DreamAudio: Customized Text-to-Audio Generation with Diffusion Models
  6. 2026Instrumental Text-to-Music Generation with Auxiliary Conditioning Branches ↗
  7. 2025ControlAudio: Tackling Text-Guided, Timing-Indicated and Intelligible Audio Generation via Progressive Diffusion Modeling
  8. 2023Noise2Music: Text-conditioned Music Generation with Diffusion Models