New field note: pick the right aspect ratio before your first text-to-video generation

Text-to-video models render at fixed internal resolutions and then crop or pad to your target aspect ratio. If you skip this step, you may get unexpected letterboxing or a subject that is off-center.
We published a hands-on note covering which ratios each model handles natively, how to preview the crop region before spending a credit, and when to choose 16:9 vs 9:16 vs 1:1 for different platforms. The article is on the blog index for all three site languages. Questions or feedback: sapsap@qq.com.