I’ve been playing a lot with the new
Photoshop AI features and today I will present the 10 useful actions you
can use in your work right now. Let’s jump right into it! We can simply add something to an
image, I think my place lacks greenery. Let’s add a plant here. Click “Generate Fill” The more descriptive you will
be with the prompt the better. Once you are done, just hit enter. Here you can choose 1 of 3 versions. I will go with this option. Also, I really wish I have 1 million subscribers, so I will replace my 100K award with the
gold creator Award for 1 million subs. Hmmm… A.I. is not there yet,
so I think it will be easier if you guys just click the subscribe button ;) We can also remove stuff from an image. Let’s say you have to photo-match
a building into the photo, but you have some distracting elements there. By the way, I made a pack with all
the images I will be using today, which you can download from my website. Link
in the corner and description below the video. Not a problem. Simply select all the elements
and click “Generate Fill”. I am using the lasso tool. Let’s skip this process. We don’t need to use any prompt in this case. And here we go, almost perfect. I probably didn’t do a great
job with the selection. Let’s select the best option. We can select imperfections,
and generate them again. This time we don’t need a prompt either. Way better. Here is before and after. A.I. is perfect for things like that. Let me quickly show you how
hard it used to be before. I am using a content-aware fill
which used to be the quickest way. Terrible job, it would
require a lot of manual work. Even the sky is not matching well. Again, here is the A.I. result. Have you got too narrow a view and
you need a little extra context? Not a problem, you can expand images. Add a space where you need to add to an image. Use a crop tool, hold ALT to expand
it in all directions at once. Select all using CTRL + A shortcut and
remove an image from the selection. Hold ALT to do so. Leave a little bit of the image
selected, so the blend is nice. Just click generate, again without prompt. Wait for it. WOW, that’s really impressive. Here are our 3 options. I like option 1 the most, but also
really like the sun glare from option 2. It’s also super impressive, in that it can figure out the sun's
position based on the original image. I will just duplicate it, change it to 1
and mask out the area where the sun is. Awesome, I really like that. We can add reflection to the existing photography.
It used to be a really hard thing to do. Here I have quickly added the house to the main
photograph with the mountain in the background. All is more or less fine, but the
reflection in the river is not accurate. Let’s add it using the Generative Fill. Let’s select the area using the lasso tool,
and click “Generate fill”, same as before. I will try without any prompt. Not great, I think we need to write something. Still not, let’s try again. A bit better, but we definitely
don’t need the ice cubes. Sometimes you need to generate a
lot of times to get a good result. Here is the best option I got. I think we need to mask it out a bit here though. As I said it would be really hard to do manually. Look well, there is potential for even
better results if I generate more options. Here is a before and after. Really like this effect here. Next, let’s replace the sky. Here I have a photo with the overexposed sky. It’s white and boring, let’s
see if we can improve it. I will select the sky. We have reflections too, let’s
see how it does with that. I will type “Sky with clouds”
in the prompt section. No, no, no. We have to try without the reflections. That one is good. I will make it more subtle though. Let’s lower the opacity. I will mask it out from here
and I think it’s pretty good. Let’s stay with this photo for a bit longer. Next, I will show you how to add
shadows to cutout people. Here I have a photo with the person I want to add. Let’s select it and make a mask. I will select her using the select subject tool. Use the “Q” key to enter into quick mask mode. Red parts are unselected. Use a white brush to add to the
selection and a black to subtract. We can just paint out the
parts we don’t want to include. Click “Q” again to exit this mode. As you can see the selection is here. We need to place her in the right place. I will place it here over this
line so it’s more challenging. Doing overcast shadows it’s very challenging
and it takes time to do it right. I will make a selection so
part of the shoe is included. We don’t need to use any prompt in this case. Wow, look at that, option 1 is amazing. It even matched the perspective a bit. The shoe is a bit different
but I don’t mind it at all. Let’s see how it does with the other one. I will subtract the top part of the shoe
from the selection so it stays the same. This might be harder because of the line. Again, the result is amazing. I think this is the way to go
with this task from now on. You can use A.I. to apply the client’s comments,
especially if they are not too complicated. Let’s say the client likes
to add water to the tap. We can easily do it using this tool. Let’s select using the lasso tool and
generate using the “running water” prompt. Here it is, easy. We can also use the power of A.I.
to improve the natural elements of an image like trees, vegetation, rocks, and so on. Let’s try to improve this tree. Select it as always and generate. We don’t need any prompt here. I don’t like any of these
options, I will do it again. Ooo, this one is pretty good. It’s very similar but looks a
bit more realistic and less 3D. I also try to improve on this
trunk but without success. First, I got this. Then I change the name to
stump, and it got a bit better. Finally, I added a palm to the prompt. It started well until I got
an image of a dead body. … I like 3D people models but they
often look fake, and you can tell it’s 3D right away. But we can improve them a lot. First, let’s replace the towel. I made a better-looking towel using
scattering, but it still can look way better. Select as always, and click generate. It violates user guidelines. I suppose it created
a naked woman and doesn’t want to show it. Let’s type the towel in the prompt. Looking great. The third option looks very well. The biggest weakness of 3D people
is hair, let’s replace that. This doesn’t look good. I included these failed tries on purpose to show you that sometimes it takes
time to get good results. I will modify the selection a bit. Now we have way better results. I will go with this one. I will do the same with the man. Let’s start with the hair. Great, and the towel. Haha, well, not what I expected. Omg. Sometimes you can get really interesting results! Let’s cover it. Great, we have a warning again. Let’s try to generate it again. Better, omg even more! The third option is fine, we
just need to cover this part. Ok, it’s a bit too bright. I am scared to generate it again though. Lastly, we can change the outfit of the person. Imagine you have a perfect cutout but
the clothing doesn’t match the scenario. For example, he should have
long jeans, not shorts. As always, make a selection. Type long jeans, and generate. Perfect, what can I say more? Let’s remove the tattoo as well, I remember
I got comments like this in the past. Let’s try without a prompt first. Nice and easy. As you can see A.I. deals with
this kind of tasks really well. If you want to learn more about visualizations or animations visit our shop or watch
more videos here on YouTube. Bye, bye.
