Is 1MP fixed resizing necessary for the TextEncodeQwenImageEdit node?

### Feature Idea

Currently, the TextEncodeQwenImageEdit node automatically rescales input images so that their total pixel count is fixed to around 1 million.
Is it technically necessary for this processing step to always rescale inputs to ~1 million pixels?

## Specific Concerns

### Resolution Range
- The base Qwen-Image model works well in the range of 1.5M–17M pixels.
- Qwen-Image-Edit also works properly within this range without particular issues.
- In fact, image quality tends to improve as resolution increases.

**Example**
`prompt : remove the human, keep film grain`

input




1024px




1328px

**Workflow**

[Qwen-Image-Edit_fp8.json](https://github.com/user-attachments/files/21922959/Qwen-Image-Edit_fp8.json)

### User Experience Issues
- Since the TextEncodeQwenImageEdit node forcibly resizes to 1 million pixels (while preserving aspect ratio), when the latent image passed to KSampler is not originally 1M pixels, unintended zooming may occur in the output.
- Users who are not aware of this internal resizing may encounter unexpected zooming, which could be frustrating.

## Summary
- Is there a specific technical reason why Qwen-Image-Edit must be fixed at 1 million pixels?
- Since existing workflow templates already include a **Scale Image to Total Pixels** node, wouldn't it be sufficient to simply explain the recommended resolution in documentation rather than enforcing it within this node?

### Existing Solutions

_No response_

### Other

_No response_