Image-Image Generation - Feeding into WAN 2.2 I2V Rapid AIO
So I want to make sure I have a realistic workflow given my specs:
9950X3D CPU
96GB System RAM
RTX 5090 GPU
-Take lower resolution input image in Qwen Image AIO, have the output generated at 2560x1440p after prompting.
-Downscale the output via Lanczos to 720p.
-Upload the output into WAN 2.2 Rapid AIO as input image.
-Set WAN 2.2 output to 720p (matching the input image).
Would this produce the best quality video? I can't really go above 720p reliably with my 5090 unless I want to deal with offload onto system memory.
Also I noticed that there is no Save Image node in the Qwen Image Rapid AIO, not sure if that is intentional or not.
"Best quality video" is subjective and basically should be whatever you find acceptable.
Feel free to add in a "Save Image" node.
Using Qwen Image Edit for video I2V (and particularly first to last frame) is a great use case.
Thanks for the quick reply Phr00t. I appreciate all of your hard work on these AIOs!