每日 arXiv

音频 · 语音 · 音乐 · 声学

每日新爬取的 arXiv 论文，按评分排序。中文摘要帮你快速判断是否值得深读。

2026-10-04
日期2026-10-04
已评分—
均分—
最高—

Daily Papers — 2026-10-04

1 papers on audio, speech, music, and acoustics.

1. Kandinsky 6.0 Video: Foundation Models for Synchronized Video and Audio Generation

Authors: Team Kandinsky, Julia Agafonova, Bulat Akhmatov, Mikhail Aksyutin, Grigorii Alekseenko et al.

Categories: cs.CV, cs.AI, cs.LG, cs.MM | Technical report on the open-source T2AV model. GitHub: https://github.com/kandinskylab/kandinsky-6 Score: 5.42/10 (Obj:8 Id:5 Ind:3 Comp:5 Eff:6 Nov:4 Rep:8)

  • Strength: 29B 全阶段权重按 MIT 开源，可在 16 GB 卡运行，人评对 LTX 2.5 的整体视觉偏好为 0.49 对 0.26。
  • Weakness: VABench 同步指标同时用于训练数据过滤和 RL 奖励；RL 的 WER −47% 由奖励模型 Whisper 测得，换用独立 ASR 后 Pro 的 WER（0.117）在三者中最差，且缺少 Ovi、MOVA 等同构开源基线。
  • Full review: Claude Code 全文七公理审稿

We present Kandinsky 6.0 Video, a family of foundation diffusion models for synchronized text-to-audio-video generation, comprising Kandinsky 6.0 Video Lite (3B parameters) and Kandinsky 6.0 Video Pro (29B parameters). Both models generate 5-second video clips with synchronized 44 kHz audio, including lip-sync, in text-to-audio-video (T2AV) and image-to-audio-video (I2AV) modes; a built-in super-resolution model raises the output resolution to Full-HD (1920$\times$1080). Building on the video generation capabilities of Kandinsky 5.0, Kandinsky 6.0 Video employs a dual-stream CrossDiT architecture that connects a pretrained video stream and a newly trained audio stream through bidirectional cross-attention for temporal and semantic alignment. Our continuous pretraining strategy first trains the audio stream from scratch on large-scale audio corpora and then trains both streams jointly on paired audio-video data while preserving unimodal fidelity; pretraining is followed by supervised fine-tuning, reinforcement-learning-based post-training, and distillation. In side-by-side human evaluation, Kandinsky 6.0 Video Pro clearly outperforms its predecessor, Kandinsky 5.0 Video Pro, and remains competitive with leading audio-video generation models, particularly in speech quality. To accelerate open research and deployment in multimedia generation, we release the code, model checkpoints, and diffusers integration under the MIT license.