This is an old tutorial series, but it was recommended to me a few years ago and it's still great. I think some buttons moved around, so if you're looking at it and confused by a button not being there say so and someone here might be able to tell you where it is.
He goes over the basics and it's done in a really good, straightforward manner. There could be better tutorials out there, but I haven't gone looking and they're probably not easily available on youtube with an easy-to-understand accent.https://www.youtube.com/watch?v=7ZlXagXwcn4&list=PL5dM8U1PBdky2jMafpYp3_xEHaSvyzBNi
I don't like that tool. Maybe I don't know how to use it, but Magic Wand seems to work a lot better for me because it offers more control.
You can use individual clicks to add to it and there's options like Tolerance, Anti-Aliasing and, a huge time saver, "contiguous" if you want to select that color all over the image instead of just a direct connection.
I'd suggest you try it. You click the tool, adjust the settings at top, then you click the desired area. You can hold shift and click another area to add to it, or hold control to subtract from it- in this way you can build up a bunch of colors all at once.
Oh well that's a bit disappointing, was hoping someone knew how to work the quick select like you see people do in TV or on youtube... As for the Wand I do use the other one quite often, but I've never tried (or knew) of the contiguous or anti-aliasing options. Guess I'll have to see how they stack up against the normal magic wand.
>>86914> was hoping someone knew how to work the quick select
Well, it's only been a few minutes...
Personally, I'm a big fan of GIMP, although I've been meaning to learn to use Photoshop for a while. The layout and seeming necessity to learn non-standard macros really messes with me every time I try to use it for more than a few minutes and then I go back to GIMP. It's really troublesome. I really need to learn to use it in particular because GIMP is really awful for drawing and so it's not at all suitable for using my wacom tablet with.>>86907
I concur with >>86908
. Using the magic wand tool and then adjusting tolerance will probably just about always be better. Alternatively, the polygonal lasso tool works very well for selecting out things by hand if you want to get perfect edges.
Having GIMP be your first image editor is kind of like learning Perl as your first programming language. On its own, it's fine for what it is, and it even excels at certain things. But it implements a good chunk of its core features in completely archaic, nonsensical ways, and if it's the only image editor you've ever used you're going to have trouble adapting to something more standard like Photoshop or Krita.
You could even say that it gimps
>>86935>But it implements a good chunk of its core features in completely archaic, nonsensical ways
Such as?... Personally, my biggest hump is just the macros that don't make sense. Like Ctrl+Z not actually being undo, but instead Ctrl+Shift+Z being undo, and then Ctrl+Z is actually redo instead of Ctrl+Y, or Ctrl+Scroll not being zoom but scroll up/down, and then scroll not being scroll but zoom. All that stuff adds up and breaks convention with just about every piece of software in existence (unless for whatever reason other image editing programs follow Photoshop's macros, which I guess would make sense?).
Oh yeah, that reminds me of one of the major reasons I don't use GIMP. Because the text feature is gimped compared to photoshop.
Does anyone know how to do those old trick images with photoshop on kissu?
Which ones? Do you mean the images that look interlaced because it looks like Image A when as a thumbnail but is actually B? Like the "Click this image to hit Alice with a shovel"? If so, I know the image you're talking about but I'm not sure I could find it.
I don't think that technique works outside of 4chan from what I've seen people test, but I've seen people come close I think... maybe?
Yeah, like one of those. Oh, guess it must be tough to pull off then
i don't know if people are aware of it but microsoft ICE, which is actually a panorama stitcher, is super good for anime stitches as well
you just feed it images (they can overlap extremely roughly and it'll still work) and it will take care of everything automatically (even stuff like changing lighting levels during panning)
it's super good
just make apngs for the sites that support it
To add to >>87470
. Basically, you just create a two frame, non-looping animation with the first frame being the thumbnail, so that when it's opened it immediately switches to the second frame. You can technically also do the same thing with a GIF, but you'll be working with a reduced color palette.>>>/ec/7281
is an example I made a while ago.
Here's the usual instructions.
Also I remember on 4/qa/ there was a thread where I was playing with related techniques.https://desuarchive.org/qa/thread/1470327/https://gist.github.com/ccd0/519b12828f0fcb4d8f4079387dc4b183
a shame that the images start looking somewhat ugly, but it's still cool
there has to be a better method, though
That's what you get for cheating on Adobe
Ah, yeah I know what you're referring to because I do it myself usually, but gifs are annoying because you need to do every frame manually. Maybe there's a script or something? Anyway, to do it in photoshop:
Filter > Other > Maximum to thin lines
Filter > Other > Minimum to thicken lines
I had googled "aquatic font" and got this "Underwater Love" one that looked good.https://www.1001fonts.com/underwater-love-font.html
This image is from before I decided to change colors, but you can see that I gave it a gradient and stroke in layer options.
I googled "hollow photoshop text" and found that you need to first rasterize the type layer...
Looks nice. I like the subtle translucency.
Lately it's all been cropping screenshots for me.
I am asking in good faith if someone could help me animate a world map to show progression in the world that I'm drawing, I'm making the individual map parts (frames) but I don't know how to do it. I don't have photoshop but I do have Gimp, (no bully) and Paint.net (I requested no bullying) so if anyone would like to help drop a line
like as a gif? That's pretty easy in Gimp.
1. Set each of your frames as a layer, in the order you want to see them (first image on the bottom, last image on the top).
2. (optional) For most efficient filesize, first go to Filters --> Animation --> Optimize (For GIF). this will generate a new file with redundant parts of each layer converted to transparency.
3. Go to File --> Export As... and save as a .gif file.
4. Save "As Animation", choose your frame delay, set frame disposal to cumulative (if you did Step 2), and click both the bottom boxes to force the chosen frame/disposal method for all frames. If you're okay with all the frames having the same delay that's all you need to do.
5. (optional) if you want different delays on frames, you'll need to reopen the file and manually edit the name of each layer. The layers will have names like "Frame 7 (100ms) (combine)", you can edit the 100ms to be whatever delay you want for that frame. Then resave as a gif like in Step 4, EXCEPT make sure to UNCHECK the box to force the default delay for all frames.
The first thing I do is duplicate the layer (it's always good to have a backup)
I make a new layer and make it pure white and position it below this active layer (#ffffff)
I lock the transparent pixels with this button.
For this I'm going to isolate the stars. So I paint them pure black. (they're semi-transparent over a white background so they look grey).
For the stuff I want to exclude I paint it pure white so it blends into the layer below it.
After I'm done painting it, I merge it with the layer below
I left click the base image to leave masking edit mode and you can see the mask is in effect. At this point you could right click the mask thumbnail to the right of the image thumbnail and hit "Apply Mask" if you want to finalize it.
I can go back into edit mask mode (Alt left click its thumbnail) and then hit Ctrl I to convert the colors, which means instead of hiding the stars I am now showing only them.
So, the background stuff has been separated into two separate layers in a non-destructive manner. It's a pretty handy thing to know how to do.
The End (?)
>>93783>Ctrl I to convert the colors
BAH! I mean INVERT. As in black becomes white.
Not necessarily a photoshop question, but I feel that it fits in with the theme of this thread. Does anyone know how to smoothly put in an image into a video without it looking out of place? Like with this video, I want to see if I can put a Miku picture onto this while in motion in a way that makes it look like an actual sticker on the car. But while trying I run into some major issues that I just don't know how to resolve. Such as making the crop proper while making it disappear with the car, actually aligning the perspective of the Miku sticker with the car and not having it completely mess up while the car turns, or having some anchor points for the perspective so it can actually "stick" to the car in a static position. Aside from that the remaining issue is that I can't figure out how to make the texture of it look like it fits in the video and is actually a part of the car.
Man, that's waaaay above me. What you already did there is already alien to me in regards to video editing. What software are you using there?
I was just using Davinci Resolve in that, but I think I might want to try Blender too going by what I've seen people make with it in the past. And as for the video it's just a pretty rough mock-up of some things that are just really bugging me that I can't figure out with the basic knowledge I have of Resolve. It's not anything too complex once you figure out how keyframing works.
>>95592>Such as making the crop proper while making it disappear with the car
I don't think there's good way other than masking individual frames manually. >actually aligning the perspective of the Miku sticker with the car and not having it completely mess up while the car turns, or having some anchor points for the perspective so it can actually "stick" to the car in a static position
You will need to use some kind of motion tracking software so that perspective is preserved no matter how the car or camera moves. This is also what serious fansubbers do for handling position and perspective of on screen texts. >Aside from that the remaining issue is that I can't figure out how to make the texture of it look like it fits in the video and is actually a part of the car.
Try to look into other image composition modes, such as multiplication.
Looks like some very cute hair there.
Unfortunately, there isn't an automatic way of dealing with that very well. It's easier these days, but for the ideal result you need to do it manually. For the temporary CSS stuff I edit for kissu I rarely bother with perfect selections because it's far too much work for something seen for a few hours and then never again.
There's two things that can help, but there's no magic fix.
(Assuming this is photoshop)
After you have magic wand selected, you can try turning on or off the Anti-Aliasing checkbox at the top that's to the right of 'Tolerance'. You generally want it to be "on", but for pixel art or other stuff that has hard edges you want it to be off. Tolerance
itself is also something you could mess with. All of this stuff has to be configured before you make the selection of the area to remove.
The other option is to edit the selection after you've made it, but it's cumbersome since you're likely working with a large, uneven selection if you're using the magic wand tool.
Make the magic wand selection and navigate to this menu under Selection
. I used 'feather' here to demonstrate that I've shrunk the magic wand selection, which you should be able to see strongly to the right.
In the end there's no perfection solution. Maybe AI stuff will be able to do it.
What does photoshop get right in its selection tools that GIMP doesn't
I guess it's similar to the reason gimp can't draw good diagonal lines or do text well at all. Whatever algorithms are involved with calculating varying degrees of opacity to make transitions between color look smooth is just not there.
(Sorry Sachiko, I forgot I was drawing on your face)
Also, the bird in the filesharing thread told me that he's still available to deliver something I shared from from a private tracker that is relevant to this discussion. >>81302
Learning a new program is awkward, but it's worth it.
Sachiko needs to ask questions if she wants to learn!
Sachiko is a girl that needs to ask questions and receive answers. You might even say she's a /qa/ girl.
Masks are a non-destructive way to remove parts of an image. I use them in the posts starting at: >>93775
It's a grayscale attachment to a layer that dictates which parts are active and visible and which are hidden. 100% black is 100% hidden while 100% white is 100% visible. If you were to look at that image's mask it would be 100% black on the statue with some grey in the transition area, which you can see where it's feathered a bit on its right side (our left). In this example image I have a Kuon layer on top of a solid green layer, and I used a black brush to mask where I drew.
The eye icon is usually to indicate whether a whole layer is active and visible or not, not its mask. But, maybe Krita does it differently.
i've done stuff like that before, got stable diffusion to generate images of blood spatters and even with the same motivation of avoiding kuso stock photo watermarks. remember when you could google image search and it would take you directly to the high resolution image without some no-name stock photo site getting in the way and shitting it up with watermarks? anyway in GIMP i extracted the blood spatter with the help of magic wand select on the white bg, feathering of the selection, then colour to alpha. GIMP's colour to alpha is a bit janky so you have to help it along with a suitable selection to prevent pixels in the middle from turning translucent
kind of strange that you aren't using ai to automatically clean background when you're already using it to generate images...
dont think stable diffusion in the webui supports alpha channels like that