Content Aware Fill for black edges

I would thinking the other day. DXO lens correction and ability to show full image with correction but also with empty edges is great. And often going to Photoshop is required to fill in those empty edges and make the image utilize maximum of what camera could capture. But I was thinking, would it be possible at some point in the future to see native DXO feature to fill in areas that are outside the corrected image, so we don’t have to go to Photoshop for it?

Microsoft Image Composite Editor (ICE) is an advanced panoramic image stitcher created by the Microsoft Research Interactive Visual Media Group. It creates a high-resolution panorama by combining a set of overlapping photographs of a scene shot from a single camera location. ICE is no longer supported but it is avilable for download.

It had a feature similar to Adobe Content Aware Fill, called Automatic image completion

ICE can now fill in any missing pixels around the edges of your panorama, making a smooth boundary even in cases where you missed a shot.

When stitching panoramas it worked as good or better than Adobe Content Aware Fill and this was many years ago. I’m sure DXO could do it in the future and it would be very useful for those black areas after lens correction that would than avoid having to crop in and lose some of the image. I think this would compliment already great features DXO offers for lens correction.

This would also be very useful to finish out the black edges of an image rotation. Especially if you didn’t adjust the crop exactly right because, you know, PL has no option to constrain the crop to the available image data.

I voted. (You should too)

2 Likes

that should be corrected by dxo as this really makes sense. you should start a thread about it so we can vote, i’m not going to start it for you =]

it was good for it’s time for sure, didn’t know you could still use it though as they let it down.

1 Like

There are 2 or three already dating back years. I’ve kind of lost hope…. Even the little stuff that is easy to fix isn’t getting fixed. This is literally a check box in the UI, a flag in the DOP/Database per image, and a little bit of logic (middle school geometry) in the code that handles the dragging of the crop window.

see →


taken from PL6.6.1 Build 199

The key is to carefully frame your shot at the time of taking. Either frame it square or allow extra room around the subject. Simples.

Oh please. Spare me the lecture. This is what I always talking about before.

Perhaps DXO should remove the rotation and cropping feature and we should get it right in camera. (Of course I’m being sarcastic)

Of course that is right in theory and I’m sure we all do that. However, in practice sometimes we find ourselves trapped against a back wall at the limit of our wide angle zoom. What are we supposed to do then? I am certainly not prepared to lug around an extra lens costing thousands of pounds/Euros/Dollars, just on the off-chance that I might need it for one shot.

We all make mistakes sometimes. I suffer from double vision, which results in my head being tilted slightly to the right. I have to concentrate hard to get my shots level and occasionally when taking a quick shot I forget.

So I absolutely agree that this is a reasonable request and not especially difficult to implement.

Great idea - yes please!!

I’m a newbie here. How do I vote for it?

I agree with MikeR Joanna.

I know there is a lot of and maybe increasing number of photographers that don´t want to postprocess their images. In that case it makes sense to use a tight framing but if you might postprocess later of any reason, using keystoning to rectify walls e.t.c. you would better leave quite a lot of space to make that possible.

I find me from time to time postprocessing old color slide images tightly framed as was common in analog times, because they were all framed without any thought at all of the future needs digital postprocessing would impose in that process.

It´s often impossible today to change what could not be changed even 40-50 years ago and it still is hard to know how a digital image really will be used during its life time too. Of that reason no simple “scout rules” will necessarily do the trick 40-50 years later no matter if the problem is just revolving around color slides or digitally born images. Even with the later I often sit there wishing I had left a little bit more space around my motif’s central parts. Too late to do anything about cases like that too.

Go to the top of the page and click on vote.

While that is a pretty standard practice for photographers, I think you are missing the point. It’s not a threat to the photography process itself, but to post production. Like the DXO’s Miniature Effect. Can’t you just get tilt-shift lens? Sure, you can. But you can also do it in postproduction. If the effect is satisfactory, why not do it in post?

It’s not a matter of mistakes in composition or the photography process, it’s a matter of maximizing the frame via software in post-production. Just as you would do so using any number of other tools. It can be done in Photoshop, but since DXO offers so many great tools for lens correction and manufacturers of lenses tend to compromise because they produce smaller lenses, lighter lenses with fewer glass elements, and ultimately cheaper lenses, it is nice to have software correction. Since we correct for so much, from chromatic aberration, pincushion and barrel distortion, vignetting, perspective, horizon line, etc., even new tools such as Miniature Effect and ReShape tools. So this seems like a reasonable next step. Technology exists in other programs, and it would be a good fit for DXO to complement their other features. It has nothing to do with where one stands to take the picture.

Considering there are Generative Fill type effects that Adobe is pushing now that can do a lot more than what I was suggesting, I would think that this older technology could be something DXO as a smaller non AI specific company could develop or outsource or license. The technology for it is 4-5 years old. Actually even Lightroom added it, except they use it for panorama stitching to fill in the blank areas. So why not add the same technology to lens correction?

1 Like

In one way I agree with @Joanna but I also agree with you. What I have found with digital photography is that it can make us lazy. Where we should be getting it right in camera, sometimes is easier to leave it for post processing, thus making us tend to get lazier and lazier and then forgetting it should be right in the camera first.

1 Like

I suppose what bugs me is that, nowadays, nobody seems interested in the craft of photography. By that I mean “drawing with light”. It seems to be a case of pointing the camera in the general direction, press the shutter and hope some digital magic somewhere can fix the total lack of forethought.

Now you’ve got the thorny subject of generative AI, which can “complete” images by referring to a massive database and inserting something that matches a description. Well, context aware fill is simply a local version of that and here, once again, people can’t even be bothered to find their own fill material, instead they rely on someone else’s idea of what might work.

It seems that some folks ought to stop calling themselves photographers and adopt the title of “digital manipulation artist”.

I’d like to see what CAF does to this…

Sorry @Joanna I nearly fell off my seat, laughing. You sure you didn’t go out specially to get that one.

2 Likes

I don’t know if that’s meant to be sarcastic or not, but she has a point.

George

Well, if you want to talk about AI ethics, cultural decay and craft and art of photography in another threat, I’ll be more than happy to. I have plenty of things to say and perhaps we would agree on many.

But since you have started the topic here. I must say as a part of unsolicited critique, if you are so interested in art of photography, why do your photographs looks so unimpressive? I don’t meat to be meanspirited here, but I’m curious. Are you really a dedicated photographer looking to improve your craft. Because you seem to be good with the technical stuff, but when I look at the images you post, they are unremarkable. So I must ask, is your frustration related to your work or culture in general?

As for digital manipulation, or analog manipulation it was always part of photography process. Especially in some genres.

Ansel Adams would be the first to tell you and he actually wrote books about it. The Camera, The Negative, The Print. Showing that its really the end results that count, not the tools you used to get there. I come from post processing world. It does not make me discard nature or value of photography. I used to do things manually as much as automated. But it really is not about the tools as much as intent.

Let me put it this way.

"The true authenticity of photographs for me is that they usually manipulate and lie about what is in front of the camera, but never lie about the intentions behind the camera.” ― Wolfgang Tillmans, German photographer

It is hard to live only from photography and unless you are in specific professions like maybe photo journalism, you will either retouch images or use the camera as capture device for photo compositing such as in commercial field or maybe fine art etc. Its about intention and integrity of the artist, not so much the tools he uses.

Here is some of my own work. I didn’t capture the images, I just worked on them. I also photograph myself, but I don’t make a living from it and I don’t think I would want to just leave it as is, even the best photograph can be improved in post production or for someone like me, its where the satisfaction is. I feel comfortable in post produciton so that is where I do my work.

1 Like

Heheheh. No, I just wanted to demonstrate what I am guessing some folks would expect software to correct. Here’s the processed original…

Someone accused the “smartphone photographers” of that and its true, that most point and shoot, as the device was made for that. Lot of snapshots not too many “photographs”. And off course lot of selfies.

But yes. As a general comment I agree that as the cameras became more powerful and automated we started to rely more of technology and less on ourselves and in the process quality of results suffers, since composition, lighting etc still work as it did before. cameras are capture devices. And if one has uninteresting mind, camera will record that. It won’t complain.

But we must not forget the people who know how and want more. One can impose a self discipline on one self and only use certain focal length, or maybe just analog camera with limited roll of film. But as long as audience is chasing new shiny pictures in attention economy of ours the new technology will eat away the market in most fields of photography. So it becomes a luxury to do it old school. Unless one can carve a niche where one can make a living of it, or one is doing it as a hobby, technology will push the means of creation and consumption of images. DXO PhotoLab is a post production tool , after all. I am asking a post production feature for that tool. Regardless how many steps back on the field I take. These requests are not mutual exclusive even if I agree with your general comment about photography these days.

Because I don’t use these fora to advertise, but to help others make their photos better using PhotoLab.

Have you seen my LF website? Of course, they are nowhere near as spectacular small - for spectacular, you need to see the printed to something like 40" x 32" or larger.

Superb work, but, from my POV, that is more graphical design than pure photography. Just not “my thing" :sunglasses:

1 Like