Skip to content
Date and source

Dates follow the news timeline in Beijing time and include both endpoints.

Oct 18Sun
  1. Reddit comfyui · 创作与工作流分享62

    Making a cinematic music video teaser with Krea 2 and MiniMax H3 in ComfyUI

    The author used Krea 2 and MiniMax H3 in ComfyUI to make a coastal noir-style music video teaser for their song No Rehearsals. The workflow has four steps: Krea 2 generates character reference sheets (close-up portrait, full-body front and back, with consistent clothing), reference-image-to-video combines the character sheets with scene or vehicle references into new shots, image-to-video is used for finalized frames, and first/last frame control handles specific changes such as a lighter being lit and both faces being revealed.

  2. Reddit comfyui · 创作与工作流分享62

    Chaining Krea 2 and MiniMax H3 in ComfyUI to make a 70-shot music video

    The author completed a 70-shot music video in ComfyUI, using Krea 2 for stills with a reference set for each character (facial sheet, body, clothing) to maintain identity consistency, and MiniMax H3 for motion in two passes: first a 0.3 MP preview to judge the motion, then continuing with the same latent to refine to 2 MP.

  3. z-ab · 开源短剧与创作项目分享17

    October 17 AI Tech Daily: Kling AI 4.0, Meta Open-Sources Muse Spark Open, and 16 Other Selected Stories

    Kling AI released version 4.0, with the competitive focus shifting from image quality improvements to director-level fine control, strengthening support for cinematic language and narrative pacing. Meta launched Muse Spark Open, an open-source multimodal model with 816 billion total parameters, supporting a 1 million token context, now available on Hugging Face. Apple released the foldable iPhone Duo and updated its App Store review guidelines to prohibit AI clothes-changing apps.

  4. Wireflow · AI短剧与视频制作教程36

    Higgsfield Ad Multiplier Alternatives: 8 Ways to Replace Ad Products or Actors (2026)

    Wireflow used the Kling O3 Edit Video node to complete an ad product replacement test: replacing a black beverage can with a terracotta mug in a 7-second 720p vertical ad, taking 3.5 minutes and consuming 142 credits (approximately $1.49–$2.13 depending on the plan). The person, action, background, and audio all remained consistent, but the mug retained the can's taller silhouette.

  5. Wireflow · AI短剧与视频制作教程42

    Higgsfield Katana Alternatives: 7 Ways to Rebuild Edits with Your Own Product (2026)

    Wireflow used its self-developed Edit Rebuild blueprint to reproduce the core capability of Higgsfield Katana: replacing a 7-second, 3-shot beverage ad with a terracotta mug, consuming 127 credits (approximately $1.33 to $1.91 depending on the plan), completed in under 5 minutes, with each edit point within 0.07 seconds of the reference video.

  6. note · koza_kura3 · 图像视频创作分享17

    Complete Your First Submission with AI Animation in 7 Days: A 1-Hour-a-Day Textbook-Free Workflow

    A 7-day production workflow for AI animation beginners, investing 1 hour per day, aimed at completing a first submittable video. The workflow uses Bing Image Creator to generate still images, Kling to convert them into approximately 3-second motion videos, and then free FFmpeg on Windows to output a 9:16 vertical silent MP4 at 720×1280px. The author's hands-on test incurred no additional cost but consumed 30 credits already in the Kling account; Midjourney, CapCut, and SNS submission were not completed in this verification.

  7. note · h_kitamura · 图像视频创作分享22

    Image and Video Generation AI News: October 17, 2026 Issue

    Several Japanese companies launched AI video production services that do not rely on live-action shooting: Media Exceed's SAKUMOVI charges 50,000 yen for a single video up to 30 seconds and 80,000 yen for up to 60 seconds; Sanctuary's One Step Studio supports 10 languages and generates a 10-minute feature in about 15 minutes; KASHIKA's Pro ai integrates 6 production functions into a single interface, and video generation can use up to 3 reference images.

  8. note · h_kitamura · 图像视频创作分享22

    Image and Video Generation AI News October 18, 2026 Issue: IMAKAI Email Newsletter Manga Adaptation, ChatGPT-Linked Manga Tool, Fujitsu 45-Second Motion Capture Footage, Qwen Transparent Background Generation Model

    Alibaba Qwen released an integrated image generation and editing model on September 20 that can directly generate and edit transparent background images, eliminating the need for background removal. It supports up to 10 reference images and outputs standard 2K (2048×2048 for 1:1). The weights are open on Hugging Face and ModelScope, but free use is limited to research or evaluation purposes; commercial use requires separate authorization.

  9. note · zou_aidouga · 图像视频创作分享58

    Using Claude Code to turn 24 AI video lighting terms into 3 explainer videos and a quick-reference table

    With the theme "light and texture in AI video prompts," the author used the Motion Explainer workflow published by Shohei Otaki together with Claude Code to create 3 explainer videos each containing 8 animations and a quick-reference table, taking about 15 minutes from listing the names to finishing the videos.

  10. note · toshia_fuji · 图像视频创作分享22

    Roundup of generative AI updates including the MOSS-Image series, DeepSeek-V4.1-Flash-1017, and the Pika X ChatGPT plugin

    OpenMOSS released the approximately 5-billion-parameter image generation model MOSS-Image-Base, supporting prompts up to 4K tokens, JSON layout control, and rendering of Chinese and English text within images. The weights are released under Apache 2.0, and Comfy-Org simultaneously launched a ComfyUI version.

  11. TelosCompass · note AI视频创作实践22

    Choosing an AI video generation model: generating 2 clips under the same conditions is enough to decide

    When using DomoAI to generate AI video from photos, you don't need to look at names or reputations first when choosing a model: first write down the one thing this scene absolutely must not fail at, then narrow the candidates to 2—model A as the one you've used, and model B as another that looks suitable—and generate 1 clip each using the same source image, the same prompt, and output conditions that are as similar as possible. Compare in the order of "the point that must not fail → specified action → production burden (waiting time, credits consumed, places needing rework)," record the reason for adoption based only on this round's results, and mark untried models as "not compared."

  12. Pika62

    Pika partners with OpenAI to launch ChatGPT plugin extension that can generate product ads within a conversation

    Pika partnered with OpenAI to launch the Pika X ChatGPT plugin extension, putting Pika's creation tools directly into ChatGPT's chat and sidebar to form a new interactive creation experience. Pika used the Product Ad feature to generate a complete ad film with AI voiceover for an armchair, all completed within ChatGPT.

  13. SuSu_酥酥22

    DeepSeek-V4.1-Flash generates cinematic video with pure code

    DeepSeek-V4.1-Flash-1017 used a single prompt to generate a 2-minute 1920×1080 60fps cinematic video purely with code, themed on the history of video games in a documentary trailer style, with music fully synthesized in code and every frame generated by code, finally rendered directly to MP4. In testing, the result is close to GPT-6-astra-high and GPT-6.1-sol-max; although weaker than Opus-5.5, end-to-end pure-code video output is already strong.

  14. ComfyUI Wiki · 工作流与开源创作分享71

    MiniMax H3 Character Swap v1.1 released, size reduced to 80 MB and reference similarity improved

    MiniMax H3 Character Swap v1.1 has been released in Alissonerdx's H3 LoRA repository; the weights went live on October 18, 2026, and are listed as the recommended version on the model card.

    Why it matters: v1.1 uses per-layer dynamic rank to cut size to one-eighth of v1.0 while improving reference similarity; the two workflows can be compared to evaluate the character-swap process.

  15. Hugging Face Daily Papers34

    Painting Change Map: using spatiotemporal change maps to guide painting-process video generation

    Researchers propose Painting Change Map, a spatiotemporal control mechanism that explicitly tracks normalized structural changes over 1,000 timesteps and can guide both forward painting and reverse "unpainting" processes. The method encodes local visual differences into RGB space to achieve stroke-level fine-grained control, outperforms existing control baselines on benchmarks, and significantly surpasses all baselines under oracle conditions. The code is open-sourced.

Oct 17Sat
  1. RenderBob · 图像视频工作流72

    Kuaishou Kling 4.0 released on October 15, with HDR, continuation, and API still pending

    Kuaishou says Kling 4.0 was released on October 15, with a single generation of up to 30 seconds, Omni Reference supporting up to 15 assets and up to 10 keyframes, and integration with Kling MCP and command-line tools.

    Why it matters: Kling 4.0 is live, but HDR, two-minute continuation, and the API are still marked as coming soon; the article also lists the parameter differences between the Flash tier and the full version.

  2. FastVideo 视频模型 · 模型发布60

    FastVideo releases FastH3-QAD-v2 video-audio model

    FastVideo releases FastH3-QAD-v2, an 8-step video and audio model trained with NVFP4 weights and quantization-aware distillation (QAD), runnable on a single Blackwell GPU (RTX 5090, RTX PRO 6000) or DGX Spark.

    Why it matters: The 8-step video and audio model lands on a single Blackwell GPU with NVFP4 quantized weights, and error data versus the BF16 teacher model is provided.

  3. Wireflow · AI短剧与视频制作教程65

    Kling Motion Control hands-on: one dance clip run through three settings

    Kling Motion Control can transfer motion from a reference video onto a character image; the author used a 5-second dance clip to drive a 3D baker character, testing three settings: 3.0 Pro with orientation video, 3.0 Standard with orientation video, and 3.0 Pro with orientation image.

    Why it matters: A hands-on comparison of the same 5-second dance clip under three settings, showing differences in time, credit consumption, and image quality, which can be used to decide between Pro and Standard.

  4. Frank Houbre · AI电影制作实战50

    Higgsfield 2026 hands-on: what one director kept and what he threw away

    Director Frank Houbre tested Higgsfield 2026: one account can call more than twenty models including Seedance 2.5, Kling 3.0, Veo 3.1, and Sora 2; Cinema Studio 4.0 is based on Seedance 2.5 and provides Director's Panel parameters such as camera, lens, and aperture, with up to 50 references and 30-second clips per generation.

  5. Curious Refuge film gallery20

    Gossip Goblin releases AI film trailer "Gods Don't Give Gifts"

    Gossip Goblin unveils the AI film trailer "Gods Don't Give Gifts," set in a world called Niiro Cradle where bodies can be discarded and consciousness becomes a commodity. Four interwoven storylines explore the horrifying consequences of humanity's attempt to transcend the flesh, with the promise of immortality ultimately giving way to questions about the value of life once the connection between body and mind is lost. With ominous narration and the warning that "the undying will envy the dead," the trailer presents transcendence as a disturbing bargain rather than a divine gift.

  6. Reddit comfyui · 创作与工作流分享15

    Request for help with MiniMax H3 REF2VA workflow for generating long videos on an RTX3060

    A user is running MiniMax H3 on an RTX3060 (12GB VRAM, 64GB RAM) to generate long videos, using the int8 convrot model and int4 encoder, with memory usage around 50GB and peaks exceeding 60GB. They use the standard ComfyUI template plus VRAM/RAM cleanup nodes, but most REF2VA workflows on Civitai rely on old custom nodes or throw errors, so they are seeking a working workflow that can support long videos with 12GB of VRAM.

  7. Hugging Face Daily Papers38

    UniAVC: unified audio-video creation via progressive training and asymmetric scheduling

    UniAVC uses a single set of model weights to achieve text-conditioned joint generation, bidirectional cross-modal generation, and instruction-guided editing without task-specific adapters. It first trains each branch on unimodal data to full generation and editing capability, then uses paired data for cross-modal correspondence; a modality-decoupled scheduler independently transforms noise schedules from a shared timestep, and directly supervises the asymmetric configuration with unimodal clean samples. This enables intent-aware editing and post-inference self-refinement, regenerating only the unsatisfactory side.

  8. 我爱计算机视觉 · 公开公众号 RSS38

    Tencent Photon x NUS Release Orbia Video World Model Evaluation Benchmark

    Tencent Photon and NUS propose the Orbia video world model evaluation benchmark, using 720 exploration and revisit cases to test 27 models, checking retention of known scenes, persistence of newly generated content, and cross-view 3D consistency. SolarWM-H3 ranks first overall with 76.32, followed by Lyra2 and EVOKE in second and third place; content persistence in long video generation is generally lower than in short videos, and the average score for forward-backward trajectories in the world model group drops from 66.21 to 40.96.

  9. Reddit StableDiffusion · 创作者与工作流分享62

    ComfyUI-SceneCut Released: Automatically Detects Video Shots, Corrects Cut Points, and Exports Frame-Accurate Clips

    The author released the ComfyUI node package ComfyUI-SceneCut, which uses OmniShotCut to find every shot in a video and label dissolves, wipes, fades, and jump cuts, then uses DINOv3 to recheck cut points to avoid false cuts from flashes, strobe, and fast pans.

  10. Reddit StableDiffusion · 创作者与工作流分享55

    PotionUI 0.0.15 and 0.0.16 Released: New X/Y Comparison Grid, Model Browser, and Image Editor

    The open-source self-hosted AI generation app PotionUI released versions 0.0.15 and 0.0.16, which can generate images, video, and audio on a local GPU, natively support models such as FLUX 3 Klein, FLUX.2, Qwen-Image (2.1), Krea-2, Z-Image, SDXL, Wan, LTX, and MiniMax H3, and can also drive a ComfyUI server and support mobile use.