Hey Man, WHY DIDN'T YOU USE CONTENT AWARE FILL???
Vložit
- čas přidán 5. 09. 2024
- Recently I posted a video demonstrating how to remove something unwanted (a garbage can), from an image using Photoshop. Many people commented asking me why I didn't use Content Aware Fill to replace the garbage can. In this video, I demonstrate why I didn't.
Check out that other video here:
• How to Remove ANYTHING...
Please follow me on Instagram:
/ anthonymorganti
Checkout my Recommended Gear List:
www.amazon.com...
If you're interested in Photoshop, Lightroom, Adobe Stock, and the Creative Cloud, you can find more info here:
bit.ly/2zwQ0nW
I use this software to record my screen:
telestream.pxf...
Unsure of how to price your photography? Check the 2019 Guide to Pricing Your Photography:
amzn.to/2S1CiU7
All links above are my affiliate links. Please read my code of ethics statement for more info about my affiliations:
onlinephotogra...
THANK YOU for watching!
Recently I posted a video demonstrating how to remove something unwanted (a garbage can), from an image using Photoshop. Many people commented asking me why I didn't use Content Aware Fill to replace the garbage can. In this video, I demonstrate why I didn't.
Check out that other video here:
czcams.com/video/wAXIqtxGf2A/video.html
Please follow me on Instagram:
instagram.com/anthonymorganti/
Checkout my Recommended Gear List:
www.amazon.com/shop/anthonymorganti
If you're interested in Photoshop, Lightroom, Adobe Stock, and the Creative Cloud, you can find more info here:
bit.ly/2zwQ0nW
I use this software to record my screen:
telestream.pxf.io/DMrW2
Unsure of how to price your photography? Check the 2019 Guide to Pricing Your Photography:
amzn.to/2S1CiU7
All links above are my affiliate links. Please read my code of ethics statement for more info about my affiliations:
onlinephotographytraining.com/code-of-ethics/
THANK YOU for watching!
Thanks for the example of how A.I. isn't a total solution. I use Affinity, and its version of content aware isn't as powerful as PS, it does work on smaller things; Far away people, lines, and join seams. Removing large objects mostly gets similar off-kilter results. One can divide up the object for content aware, but if I'm going to do that, I'll just feather a hole and dup the layer and try sliding the dup around to make a patch. Doesn't work all the time, but I tend to get better results trying the patch first.
My content aware is not working..when i click to ok then shows not enough memory
Excellent example of the need to master more than ‘one fix all” tool.
Really appreciate the work you put into these tutorials Anthony.
I would have moved the bin before taking the picture, then replaced it. :)
Exactly what i was thinking...
0:29 "Hey, why didn't you use Content Aware, Phil?" I wouldn't have said that. I would have said, "Hey, why didn't you use Content Aware, Anthony?"🤣
I admit it, I just finished watching that other video and couldn't imagine why you didn't use Content Aware Fill. I would have used it on the can left of the ladder, separate from the parts between the ladder rungs where I probably would have just cloned them out with a few quick stamps. Quick selection and I don't see eye to eye. When I to try to select some itty-bitty part like the can between the rungs, it will expand the selection to include the entire sky or something, and then when I alt or option it back it will shrink to a single pixel. And I'm pretty sure I've seen it stick out its tongue at me. The selection tool mocks me, Anthony.
Now I'll watch the rest of this video and see if I'll need to eat crow, which I'm sure I will. You sound pretty confident, almost cocky like I've never heard you before.
~Sigh.~
Okay, I watched your damn video. Do you have a good crow recipe, Anthony? Prepared properly, I hear it tastes just like chicken.
In all seriousness, great pair of videos. I'm glad they showed up in my recommended feed. I've not run into a problem like this before, but if I had I'd probably have gotten pretty frustrated. Now I'll know better, and I'll do it the old-fashioned way. A question though: Couldn't you do it in two steps, the first being a selection of half the can at water level and up, and the second step a selection at sand level and below? I wonder if Content Aware could handle it then.
When you were given the green area on the auto content aware fill, you should’ve erased the green areas that were NOT conducive to helping the new image - namely any part of the ladder/wood, sky and clouds. In short, the only area left in green should’ve been the sand and the water. Try that and I think you’ll have pretty good results.
You CAN use (classic) content aware fill.
Divide the image in two by selecting precisely along the left side of the ladder with the polygon lasso tool to the edges and round the left part of the image and copy the selection to a new layer (Ctrl + J). Then make a selection from the new layer (Ctrl + left click on the layer), activate the background layer and create the inverted mask from the selection. Switch off the background layer and activate the new layer. Now with the same polygon lasso tool select (top-down) precisely along the pixel part and the transparent part from a bit over the garbage can (about half the height of the black top) and to a bit longer below the can, then diagonally (left-up) through the tracks in the sand to a bit below the water, and from there vertically to the imaginary horizontal line through your starting point, and finally right to join that point.
Invoke the content aware fill (Shift + Backspace), and you should get a fairly correct and usable result. It might be necessary to work a little with the three leftmost lines around the selected area to make the fill look perfect, and a bit of cloning may be necessary to make the lines in the sand look right. You can use the same kind of cut out for each of the openings showing the smaller parts of the garbage can, or you can use the clone tool for those smaller areas. If necessary you may want to adjust the new layer slightly with the transform tool (Ctrl + T). The idea is to prevent the content aware tool from "seeing" the parts of the image that are unwanted for the fill.
HI Anthony.. I watched your tutorial on removing the rubbish bin and I enjoyed how you made it happen..I don't have the skills yet to be able to apply your teaching within Photoshop... Why is there so much negative feedback on this subject from those that must have seen the same tutorial ???? You showed us something and I follow as best I can.... Keep up the good work Anthony.. Neville J.... West. Australia...
Your videos are terrific! Easy to understand and very informative. Thank you so much!!
Watched both videos, was able to fix my photo by following along with the first one. Not sure if this would've helped since it is an amination image. I keep both videos just in case for future needs. Thanx!
As you know the patch tool is another option, but it does not work well when the selection is up against a surface such as the wood ladder in your image - the way around that is to use the clone stamp along the the edge of the ladder so that the area being patched is separate from the ladder and then use the patch tool to replace the rest. Of course by the time you do that you might as well use the clone stamp for everything just as you did in the first place. However, the patch tool does do a pretty good job of blending the edges of the selection with the surrounding area so it may save some of that manual fussing around cleaning up the edges.
Interesting video. I was wondering if ON1 does a better job
Yup, content aware has its uses and is improving with each update but ultimately you are often better off cloning an area or using the patch tool as per your previous video. It's never good at replacing areas which are physically attached to the one you want to keep unaltered.
Why do Americans pronounce "didn't" as "dittent"?
Easiest way i found to remove the bin in tjhis instance was to make a copy of pic..then on top layer copy make selection of bin ...then use selection to copy area to left of bin...copy and past over bin merge with top copy then use eraser to delete any unwanted pixels...took me a couple of mins at most..i know theres a lot ways to skin a cat but i found that the easiest for this pic
Anthony how about doing with content aware fill but doing it in three sections? When the green comes up after you select content aware you can erase the parts you don’t want PS to copy from?
Yes Lenie. That’s the technique I normally use and I would clone stamp the bits appearing between the steps.
I knew this would happen; CAF is only really useful in similar areas or those so disparate you won't notice anomalies! With an example such as this, I just bite (byte :0) ) the bullet and use a clone stamp.
What If you select a bigger area?
How about taking it over to Luminar 4??
Cool
Thank you, great to know why! ;) and great work with both videos, thank you!
Well Done 👏 I hope they will now understand the difference
I wouldn't either use content aware fill. I would have moved that trash can 3 meters to the left before taking the picture. That way you don't have to pay adobe a lot of money for subscribing to a program that should be a one time expence. Hence no need to even use photoshop.
That, of course, assumes that you can move the trashcan. Trashcans around these parts are often tethered to an anchor so they can't be stolen.
Great comparison. I thought your first video did great. Also, since you are comparing, any chance of using Affinity Photo options on this same beach photo?
I just screen grabbed this and used the inpainting tool in Affinity. Not perfect, but 5 seconds mostly done and much better than CaF in 8 minutes. Little cloning, and job done in a couple of minutes.
Curious how this particular photo would work in Luminar
That was painful to watch. This tool, I am almost sure, is capable to clone perfectly the left side of the bin. When you open the content aware tool box, just delete the green over the bits of the image you don’t want to clone from. In this case, at least, you shouldn’t try to define the area yourself. That seems to mess the AI up. For the right side, between the steps, I would probably try the same method, or simply clone it, as it is smaller and less visible. Can this photo be downloaded somewhere to try ?
O
My comment is why take this photo in the first place?