It seems all the other RAW editing contenders are years ahead with their healing tools and brushes, yet, DxO is lagging far behind. I would honestly be interested in knowing when we might expect a competitive healing tool.
The lack of a decent healing tool is one of the primary reasons I do most of my editing in Lightroom and barely pay attention to DxO anymore. Otherwise, I’d likely be all in with PhotoLab. I’m sure I’m not alone.
PhotoLab will get one eventually. I have tried the generative erase tools in many of PhotoLab’s competitors and the majority of them have significant flaws and significant limitations. Until PhotoLab can create one without the types of serious flaws I see in their competition, I will happily continue to use their manual tools which at least puts me in full control of what will replace the objects I want removed .
It does take a lot of experience and effort to understand and master PhotoLab’s repair tool, but with that experience and expertise there is almost nothing that any of the generative tools can accomplish that can’t also be done with it. Once that expertise is gained it also gives consistently better results compared to most of the generative erase tools available forr those willing to take the time and effort.
As a general rule, generative erase tools make the decision regarding what will replace the object to be removed. For my photographs I want to be in charge of that decision.
Generative erase tools often struggle when the removing items with very complex , highly detailed, backgrounds
Mark
I sure hope so. But, honestly, Lightroom’s Generative Erase is exceptional anymore. I have used it on very complex items, where the others fail, and it generates an acceptable replacement nearly every time. Only on the rarest of occasions has it failed me and in those instances, simply brushing a smaller area made a world of difference. Luminar Neo is second on the list, though it struggles a bit more does all of its AI locally. On1 Photo RAW is still not even close to either based on my most recent use (within the past 3 months).
While I have no personal experience with Lightroom’s regenerative erase I am sure it’s probably among the best. Many of the others, like ON1’s version, are problematic at best.
Consider the size of Adobe and the resources it has. So yes, DxO has some catching up to do, but I’m hoping that whenever they implement a regenerative erase tool it will be far superior to the messy ON1 Photo Raw and similar implementations.
Mark
A skilled operator would find any of the generative ones laughable. My brother was doing straight cloning to a ridiculous degree (at clients’ behest) in plain old Photoshop over a decade ago. He reckons none of the current lot can do as good a job a lot of the time.
Then again, we’re not all someone like my brother with decades of experience. But also… I personally find it pretty rare that I need to remove something that isn’t achievable using my limited experience on the clone tool.
I’m learning all the time, too. Recently I saw how removing a branch from a bird photo is much easier if you first erase small sections to cut it into smaller pieces, then remove those. The license plate example was a classic case of wanting the tool to “do everything” but it was relatively simple to just treat it in smaller parts. Another tip is to clone away the bulk and then repair any artefacts that remain.
AI masks were a long time coming and, in general, I find them to be a HUGE time saver (and in the latest release, generally very accurate). The main difference it has made for me is that I will mask more often because of this ease. That’s not to say I never used Control Points or even a simple brush in the past, because I absolutely did when it was worth my time,
I expect “full” healing will come, but it will take time to arrive and time to mature.
I see user requirements as something of a bell curve when it comes to healing/cloning tools.
Many of us don’t need “the absolute best tool that will stand up to scrutiny when printed on a giant billboard”.
We do want “doesn’t have obvious repetition or artefacts that look out of place and require further fixing, and provides high enough fidelity that the image doesn’t become a blurry mess, even on when posted online”.
So it’s all about being good enough to work 95% of the time to an acceptable degree, rather than being absolutely flawless every time (which would be great, but unrealistic) or needs frequent further correction to the point it’s not really fit for purpose.
I cannot disagree with that, but we first need to define what it will do well 95% of the time.
If I take a photo of a crowd and ask it to remove all of the people? If I ask it to remove one of the people? If I ask it to remove a person who is the only one visible, in front of a colourful mural? A person who is standing in front of a fence? A bird that is flying across the foreground of a landscape? Or a dust spot in a clear blue sky.
There is a HUGE range of “just remove this” possibility. PhotoLab will already do 95% on the dust spot in the sky.
Take away persons? Political filtering? 
I’d be more than happy if captures of old, scratched negatives could be cleaned up with a simple possibility to review and remove fixes.
That makes sense too - different use cases require different levels of functionality. I guess it depends if we’re talking about generative AI removal or more traditional spot removal “remove this branch” or “remove this bird from the background” vs “remove this crowd”.
For me it’s a question of what tool to use and when. I use DxO when I want to quality RAW processing and lens correction to get the best results from the light and my chosen composition - tinkering with actual content is for a different tool, in my case for historical reasons, Corel Paintshop Pro
Google Gemini Nano Banana is crazy. Even in incredibly complex situations, it is completely flawless.
These tools are just like the ones used ten years ago !
I wouldn’t call “generative erase” a healing tool, personally. It’s just extra AI features to change a photo.
If DXO start adding more AI into PL then the system requirements go up and up, and then you need to buy credits to use the extra AI, more people complain, more things stop working properly… I imagine there is a strong reason for why DXO didn’t implement AI masking when everyone else was, and waited a year or 2.
That said, I find PL9 pretty bad for repairing scans of old photos. I don’t want AI to recreate them, I just want to remove blemishes, dust, scratches and sharpen up the image a little - I don’t find PL9 very intuitive or quick for that, so use my old copy of ON1 Photoraw - even without using the AI features, the repair and heal tools seem to work more accurately and quicker for me.
I realise PL9 ISN’T a photo repair tool, it’s a RAW editor, but some extra photo healing (without AI generation) would be very welcome.
I certainly hope not, as:
(A) That would definitely push me away from the product. I dislike any software that issues paying customers with limited “credits” to use its functionality. Imagine if BMW only let you use heated seats 10 times a month unless you purchase additional credits…
(B) AFAIK AI editing in PhotoLab is performed locally, so no extra straight is being put on DXO’s infrastructure, only our own systems.
Of course - at the moment, AI in PL9 is done locally.
But the features being suggested above are far more AI intensive - that pushes up the graphics card requirements and/or involves online credits. Better quality AI replacement/generation usually involves going to an online service - local services (ie on your computer) tend to be lower res. I would much prefer it if DXO doesn’t go down that route and sticks to being an excellent RAW editor. But I would certainly like better repair tools.
That’s already the reality for long time in automotive industry.
A friend had a Mercedes E220 20 years ago and the difference between E200 and E220 engines was a setting to get 30 HP more !
And Tesla Autopilot is activated for 10000 bucks !
Well that’s just it… an instant “turn off” for me. I’m old enough (and I’m not saying I’m alone here) to remember PC games coming on CD’s and DVD’s that we actually own. I am not a fan of perpetually needing licenses for software to function.
I’ve got you beat. 
I remember 45s and 33 1/3 records. I have no problem with keeping software current (not to mention the 1 TB of cloud storage I get — all for under $13 a month; compared to other storage providers, that’s practically like keeping Lightroom up to date for free). As for AI credits, since you mention it, I have yet to come even close to using up my monthly allocation - maybe I just don’t need to do a lot of editing, but the editing I do is very detailed and DxO doesn’t handle details. Frankly, it’s the only thing missing from Photo Lab that keeps me from going all in after all these years.
People get snobby about the Adobe subscription model but when you start to compare what you actually get, compared to what you don’t with DxO, you’re making a much better investment subscribing to something that continues to improve over something that is like taking a slow boat to China (and I say that after having had DxO products for nearly 2 decades).
I hear you, if you compare the cost of a year’s Adobe subscription vs. buying the PhotoLab suite, you do get more with the Adobe suite and IIRC it’s slightly cheaper too if you buy from the right places (!)
But there I’d say it depends what you use the products for. I never used the Adobe cloud storage for instance, so while it’s nice to have, it’s not something I’d factor in to my dilemma.
(And I still don’t like the idea of not owning things I’m paying for).
The one in Affinity is good In my opinion , sometime put the image on ther to finish it off.