Improve Denoise AI workflow: automatically re-order adjustments to speed up processing

[salvoo76993234]

I would like to suggest an important improvement to the Denoise AI workflow in Camera Raw and Lightroom.

Currently, if Denoise AI is applied after other edits (curves, masks, color adjustments, lens corrections, etc.), the entire pipeline is recalculated from the RAW file, resulting in significantly slower processing times—especially with high-resolution files (50–60 MP).

A manual workaround already exists and works very well:
1. Copy all current adjustments.
2. Reset the RAW to its original state.
3. Apply Denoise AI (much faster this way).
4. Paste back the adjustments.

This produces identical results but completes in a fraction of the time.
It would be extremely helpful if Camera Raw / Lightroom could perform this optimization automatically. For example, the software could:

- Temporarily store the user’s adjustments
- Apply Denoise AI on a clean RAW
- Reapply the adjustments on the new DNG
- Automatically refresh masks and profiles

This would massively improve the user experience, especially for photographers working with large RAW files and complex masks.

Thank you for considering this.

[VishalKa5]

Hi @salvoo76993234 ,

The current behavior is slow because Denoise AI is executed after all previous edits, forcing Camera Raw / Lightroom to recalculate the entire processing pipeline from the original RAW file. With large files and complex masks, this becomes very time-consuming.

In practice, Denoise AI does not depend on most user adjustments (curves, color, masks, etc.). When it is applied before any edits, processing is much faster and the final image result is the same.

A simple logical improvement would be for the software to:

• Temporarily save existing edits
• Apply Denoise AI first on a clean RAW
• Create the denoised DNG
• Reapply all edits automatically

This matches what advanced users already do manually, but without extra steps.
Implementing this order automatically would significantly improve performance without changing image quality.

Thanks,

Vishal