mlsanityData to 7 Oct 2026

Post-Training Quantization for Audio Diffusion Transformers

arXiv 2510.00313 ↗First posted 30 Sep 2025eess.AS cs.SDcited by 2

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

Abstract (arXiv)

Diffusion Transformers (DiTs) enable high-quality audio synthesis but are often computationally intensive and require substantial storage, which limits their practical deployment. In this paper, we present a comprehensive evaluation of post-training quantization (PTQ) techniques for audio DiTs, analyzing the trade-offs between static and dynamic quantization schemes. We explore two practical extensions (1) a denoising-timestep-aware smoothing method that adapts quantization scales per-input-channel and timestep to mitigate activation outliers, and (2) a lightweight low-rank adapter (LoRA)-based branch derived from singular value decomposition (SVD) to compensate for residual weight errors. Using Stable Audio Open we benchmark W8A8 and W4A8 configurations across objective metrics and human perceptual ratings. Our results show that dynamic quantization preserves fidelity even at lower precision, while static methods remain competitive with lower latency. Overall, our findings show that low-precision DiTs can retain high-fidelity generation while reducing memory usage by up to 79%.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Placed in Audio, speech and music, one of 14 directions in Diffusion models. Builds on 27 of its 30 references that are in the map. First posted on arXiv 30 Sep 2025.

Builds on

27 of 30 references.

Closest ancestors · 6

Same direction, newest first

  1. 2024ImmerseDiffusion: A Generative Spatial Audio Latent Diffusion Model19 Oct
  2. 2024PTQ4ADM: Post-Training Quantization for Efficient Text Conditional Audio Diffusion Models20 Sep
  3. 2023EDMSound: Spectrogram Based Diffusion Models for Efficient and High-Quality Audio Synthesis15 Nov
  4. 2023Controllable Music Production with Diffusion Models and Guidance Gradients1 Nov
  5. 2023Noise2Music: Text-conditioned Music Generation with Diffusion Models8 Feb
  6. 2023AudioLDM: Text-to-Audio Generation with Latent Diffusion Models29 Jan

Other references · 21

  1. 2017Attention Is All You Needcited by 59,932
  2. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  3. 2022Scalable Diffusion Models with Transformerscited by 5,931
  4. 2022SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Modelscited by 1,338
  5. 2022All are Worth Words: A ViT Backbone for Diffusion Modelscited by 498
  6. 2023DeepCache: Accelerating Diffusion Models for Freecited by 346
  7. 2023Q-Diffusion: Quantizing Diffusion Modelscited by 266
  8. 2022Post-training Quantization on Diffusion Modelscited by 262
  9. 2024Stable Audio Opencited by 192
  10. 2024Fast Timing-Conditioned Latent Audio Diffusioncited by 178
  11. 2023PTQD: Accurate Post-Training Quantization for Diffusion Modelscited by 167
  12. 2024SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Modelscited by 161
9 more
  1. 2024PTQ4DiT: Post-training Quantization for Diffusion Transformerscited by 77
  2. 2023Towards Accurate Post-training Quantization for Diffusion Modelscited by 46
  3. 2022Post-training Quantization for Neural Networks with Provable Guaranteescited by 32
  4. 2024Token Caching for Diffusion Transformer Accelerationcited by 29
  5. 2025AudioX: A Unified Framework for Anything-to-Audio Generationcited by 25
  6. 2024Sketch2Sound: Controllable Audio Generation via Time-Varying Signals and Sonic Imitationscited by 23
  7. 2024DiTAS: Quantizing Diffusion Transformers via Enhanced Activation Smoothingcited by 13
  8. 2024DEX-TTS: Diffusion-based EXpressive Text-to-Speech with Style Modeling on Time Variabilitycited by 5
  9. 2025Post-Training Quantization for Diffusion Transformer via Hierarchical Timestep Groupingcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 1; 2026: 1.

2 papers.

  1. 2025Quantizing Whisper-small: How design choices affect ASR performancecited by 0
  2. 2026Same Bit Width, Different Outcomes: Post-Training Quantization of Text-to-Speech Across Architecturescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Audio Palette: A Diffusion Transformer with Multi-Signal Conditioning for Controllable Foley Synthesis
  2. 2025LRQ-DiT: Log-Rotation Post-Training Quantization of Diffusion Transformers for Image and Video Generation
  3. 2025DiTSE: High-Fidelity Generative Speech Enhancement via Latent Diffusion Transformers
  4. 2024Q-DiT: Accurate Post-Training Quantization for Diffusion Transformers
  5. 2024Autoregressive Diffusion Transformer for Text-to-Speech Synthesis
  6. 2023From Discrete Tokens to High-Fidelity Audio Using Multi-Band Diffusion
  7. 2020DiffWave: A Versatile Diffusion Model for Audio Synthesis
  8. 2024TaQ-DiT: Time-aware Quantization for Diffusion Transformers