Change how Repair in the Healing Brush works

I’d like to suggest/request a change to the Repair function of the Healing Brush in PhotoLab. Clone and Repair are almost exactly the same. Both rely on grabbing a patch from another part of the photo to cover up a spot or area to be fixed/modified.

Clone works the same way as a clone tool in any other photo editor, so that’s fine. But if you want to see how a really good repair or healing tool works, look at the Inpainting Brush in Affinity Photo. It samples from pixels right around the area you mark with the brush. It does an excellent job and PhotoLab cannot come close to how good a job it does in many cases. I know this from working on restoring old photos. I currently have 80 photos to go on a 320 photo project.

Yes, DxO’s Repair and Clone both grab a patch from another part of the image - but Repair seems to work by blending the patch in with the selected area. It isn’t the same as clone with a reduced opacity - it seems to follow its own rules. I’m very often impressed with the result. It can be seamless and convincing. Not at all impressed with attempts to repair at the edges of the image, though - those tend to be terrible. Although DxO is well-aware of the problem, I haven’t seen any indication of what they intend to do about it.

While I haven’t tried Affinity’s approach to repair (I have tried some of AP’s other tools), it doesn’t sound like it would work well for me much of the time. I am one who likes DxO’s approach. And with the latest PhotoLab there are many adjustments that can be made to how DxO applies the source patch to the destination target.

Do you have a link to a video (or your own video that you can post here) showing how much more effective AP’s repair tool can be?

I have not made a video of myself using Affinity’s Healing and Inpainting brushes, but other people have done videos of this and put them on YouTube.

I agree that many times I can get the same result with PhotoLabs’ tools as I can with Affinity’s, especially when there are not too many spots to fix. However, I’ve found I have to use Affinity’s touch up and repair tools when there are numerous spots and scratches, because it either takes way too long and/or I cannot get the same quality result using PL’s tools.

And when I say numerous, I mean it looks like someone was holding soot in their hand and blew some of it onto the photos. So there are many spots of various sizes and shapes. Believe me, I’ve tried using PL’s tools to fix these but finally had to give up, export a tiff, and clean up the spots in Affinity.

I recently removed well over 100 spots from a photo captured by someone with an extremely dirty sensor using PL 6’s repair tool. Easy work that only took me a few minutes. How big and varied were these dirt spots. Would you be willing to share the “before” version of the image to which you were referring?

Mark

Share the image? It isn’t just one image. There are over 300 in this project and since they belong to a client, no, I cannot share them on the Internet. Some are definitely worse than others. Some I’m able to handle entirely in PL6 but I have to resort to Affinty Photo for many others, as I’ve stated.

100 spots is practically nothing on some of these images. Try 1500 to 2000 spots plus a bunch of smears, some scratches, and uneven tones due to handling (skin oils) on a single image. In some cases resortation of facial detail is needed. Imagine 4-6 hours restoring a single image.These are all from very old slides, 1950’s to 1990’s, and I snapped pics of each one in RAW and jpeg as they were projected onto a screen.

Now, to get back to the point of my original post which was not to set up a challenge to see if another PL user could efficiently remove a handful of spots from a photo in PL, I would like for the DxO programmers to look at Affinity’s Healing and Inpainting brushes to see what could be improved in PhotoLab.

I’m all for expanding the capabilities of these tools to make them more versatile. So I’ll add my vote.

I don’t have any specific information, except that there are many patents on these technologies.

Pascal