please help regarding LTX2 I2V and this weird glitchy blurryness
Users struggle with LTX2's image-to-video generation, citing weird artifacts despite following official workflows.
A viral discussion among AI creators centers on persistent technical issues with Lightricks' LTX2, a leading image-to-video (I2V) generation model. Users report that outputs are plagued by "weird glitchy blurryness" and artifacts, even when meticulously following the official workflow hosted on RunningHub.ai. Troubleshooting attempts have included using high-quality source images up to 2048x2048 resolution, experimenting with frame rates (25/48 fps), testing different samplers like LCM, adjusting the detailer setting to 0, and toggling camera LoRAs—all without resolving the core visual defects. The problem, shared by user DotNo157, suggests the issues may be inherent to the current model iteration or its hosted implementation, impacting creators who rely on it for coherent video content.
- LTX2 users encounter persistent glitchy and blurry artifacts in generated videos, despite following official workflows.
- Extensive troubleshooting—including 2048x2048 source images, 25/48 fps settings, and LCM samplers—has failed to resolve the visual defects.
- The issue highlights the unreliability of current AI video generation models for professional, artifact-free content creation.
Why It Matters
Widespread model instability hinders professional adoption of AI for video generation, slowing creative workflows and commercial projects.