Importing an iOS photo with HDR gainmap and activating HDR edit-mode causes the SDR grade to drift

[nnhuy]

The following applies (at least) to RAW/ProRAW/HEIFs produced on iPhones.

Current behavior:
1. Import a .dng/.heic with HDR gainmap into Lightroom and observe that the initial SDR preview is as expected
2. Activate "HDR" mode and see the gainmap is immediately applied as expected
3. Activate "Preview for SDR" and notice how the preview is entirely different than anything you've seen before this point in either Lightroom or iOS's native Photos app.
4. Exporting this photo will propagate the new unexpected SDR grade as the base image.

Expected behavior:
The preview outputs from pre-HDR-activation and from post-HDR-activation with "SDR" preview on are broadly similar, especially before any editts are made. Exporting an HDR image when no light/color edits have been made will also maintain the same SDR grade as the imported version.

My usecase is sometimes using Lightroom for quick denoise/sharpening/crop/perspective edits, and my expectation is that those interactions do not affect grades.

[MadManChan2000]

Thank you for reporting this and for the feature request. Your use case (just transform & detail-based edits) makes sense.

One of the challenges with this proposal is providing some measure of "editing continuity" when tone and/or color edits start to affect the result. For instance, we're trying to avoid a sudden visual jump in the SDR result when nudging Contrast from 0 to +1 (due to switching from the embedded SDR rendition to one that is derived from the HDR rendition).

In any case, I've noted your request and we'll try to find ways to make the SDR/base rendition experience better in the future, for use cases such as the ones you've described here.

[nnhuy]

Thank you for considering my point, and I look forward to seeing how the HDR editing experience evolves.

I lastly want to point out that between base SDR, gainmap HDR, and Lightroom tonemapped HDR, the first 2 have creative intent, while the last one is artificially generated. I do realize there are workflow complexities, like you mentioned, and the need for a tonemap representation to exist, as not all inputs and outputs have gainmap support. However, it still philosphically feels wrongs to force discard user-created base SDR "art" in favor of a algorithmic construct when importing a gainmap file.