The Ultimate Guide to Documentary Animations: Napoleon Project Breakdown
A four-hour, click-by-click rebuild of a painted Napoleon parallax animation, from AI upscaling in Photoshop to a sword fly-through transition in After Effects.
Posted
1 weeks ago
Duration
Format
Tutorial
educational
Views
185.5K
4K likes
57 · 43
Big Idea
The argument in one line.
A painting comes alive in documentary animation when every layer is cut cleanly, textured on its own, rigged at the real joints and moved by stacked nulls, so the camera move sells depth instead of exposing the cutouts.
Who This Is For
Read if. Skip if.
READ IF YOU ARE…
You make history or documentary YouTube videos and want your still paintings to move like the parallax shots on big explainer channels.
You already know your way around After Effects and Photoshop and want a complete real project to follow, not a feature tour.
You've tried a 2.5D parallax before and it looked like flat cardboard cutouts sliding past each other.
You want to know where AI tools (upscaling, layer decomposition, generative fill) actually help in this pipeline and where they still cost you quality.
You rig characters in DUIK and keep fighting broken elbows and messy controller setups.
SKIP IF…
You've never opened After Effects. This assumes you know layers, keyframes, the graph editor and parenting.
You want a fast overview. This is four hours of real-time work with very little summarizing.
You're after a one-click AI animation tool. The whole point here is manual control over every layer.
TL;DR
The full version, fast.
Documentary-style painting animation is a layering problem first and an animation problem second. The workflow takes three public-domain paintings, upscales and extends them to 3840x2160 with AI (testing several models and compositing the best parts), then separates each into foreground, midground and background layers by hand in Photoshop because AI layer decomposition caps at 2K and regenerates the image. Texture is applied to each layer, not as one overlay on top, so it moves with the camera. In After Effects the character is cut at the joints, auto-rigged with DUIK Angela, given looping blinks and a wiggle head, then placed in 3D space with an orbit null and animated with overlapping stacked nulls at 60fps. Scenes connect through a track-matted sword fly-through and a luma-matted fire transition.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
The finished animation plays, then a recap of every stage covered and where the starter file lives.
01:37 – 02:19
02 · The story behind the paintings
Napoleon crossing the Alps before Marengo, told through three paintings: Benoist's First Consul, David's heroic horse and Delaroche's realistic mule.
02:19 – 03:46
03 · Upscale and extend the paintings
Every painting is pushed to 3840x2160. Nano Banana Pro underperformed, GPT Image 2 did better, and the final images are composited from several generations. A top-down floor is generated for scene one.
03:46 – 04:15
04 · Layer separation: AI vs manual
Two ways to split a painting into foreground, midground and background layers.
04:15 – 05:46
05 · AI separation with Higgsfield
Higgsfield's layer decomposition splits an image into eight layers for 24 credits, but only at 2K, with no control over splits and a regenerated image.
05:46 – 27:57
06 · Manual separation in Photoshop
The select subject, duplicate, mask, Remove, merge loop repeated object by object, with clone stamp cleanup and a backup PSD for every image.
27:57 – 45:03
07 · Texturing in Photoshop
Why a single overlay texture looks fake, then paint, wood and halftone textures tuned with levels and curves and clipped to each layer before merging.
45:03 – 1:21:03
08 · Character: separate the body parts
Napoleon is cut into head, upper arms, forearms, wrists and hands, with every hidden area painted back in. The horse and rider are separated for wind animation.
1:21:03 – 1:29:56
09 · Organize and import into After Effects
An organizer script sorts the project, a custom script imports every layered PSD at once, eyelids are taken from a closed-eye image and limbs are renamed and parented.
1:29:56 – 1:34:04
10 · Rig with DUIK Angela
Anchor points move to the joints and DUIK's auto-rig builds a one-controller arm. Bad anchors break the rig; a stretched upper arm fixes the elbow.
1:34:04 – 1:55:30
11 · Animate the eyes and head
A double-masked head layer sandwiches the eyelids for blinks, looped every 7 seconds with time remapping. A gradient shape-layer neck makes a wiggle head possible.
1:55:30 – 2:19:27
12 · Animate the hands
Controller keyframes block the pointing hand and the hand sliding into the coat, smoothed with Bezier motion paths, a front coat layer and drop shadows.
2:19:27 – 2:20:46
13 · 3D camera: the basics
One-node vs two-node cameras, what focal length does to parallax, stacked overlapping nulls and why the work is rendered at 60fps.
2:20:46 – 2:44:25
14 · Place layers in 3D space
An orbit null scales layers back in depth without changing the frame. A corner-pinned floor, fake reflections, a masked table shadow and Blender-rendered birds finish the set.
2:44:25 – 2:55:30
15 · Camera animation
A 44mm camera on a chain of five overlapping nulls: birds in focus, push to the character, close-up, follow the hand, pull back to reveal the sword.
2:55:30 – 3:22:24
16 · The sword transition
A textured sword arcs over the table on a Bezier path, then the camera flies through it into scene two using a track matte, a rotating null and a CC Lens warp.
3:22:24 – 4:17:33
17 · Napoleon on the horse and the mule
Breeze wind on hair and cloak, wiggle heads, a dolly zoom, a luma-matted fire transition into the mule scene with snow, plus fixes found in preview and the text box build.
4:17:33 – 4:18:46
18 · Final render
Wrap-up and where to get the project files.
Atomic Insights
Lines worth screenshotting.
A texture layer set to Overlay on top of the finished animation stays still while the camera moves, which reads as a screen filter, not paint.
Real paint texture is embedded in the image, so it has to be clipped to each separated layer before animation, not added at the end.
Texture goes on after layer separation, because textured images give worse generative fill results when you remove objects.
No single AI model is best at extending paintings; the reliable approach is to generate around six versions and composite the sky from one and the ground from another.
GPT Image 2 cost 14 credits per image versus 4 for Nano Banana Pro, but gave the better painting extensions on this project.
AI layer decomposition tools regenerate the whole image, so you lose detail and get small variations from the original.
Higgsfield layer decomposition only works up to 2K, which rules it out for 4K documentary work.
Photoshop's loop of select subject, duplicate, add mask to selection, disable mask and Remove separates a painting into clean layers one object at a time.
Split the background into two depth layers when distant buildings and nearby trees sit far apart, or the parallax reads flat.
DUIK Angela's auto-rig works with a single controller only when the arm layers are named, parented wrist to forearm to upper arm, and anchored at the real joints.
Painted characters have no neck under the head, so a head wiggle needs a painted or shape-layer neck to hide the gap.
A Negative mask expansion of about -2 on the subtract layer closes the see-through gaps around animated eyelids.
Never animate the After Effects camera directly; parent it to a null, then chain overlapping nulls so each move starts before the previous one ends.
Slow camera moves at 24fps show motion judder, which is why painterly parallax work is rendered at 60fps despite longer render times.
A short focal length like 24mm exaggerates parallax from small moves, while 85 to 100mm compresses the scene and reduces layer separation.
A painting that shows feet needs a floor plane rotated -90 on X, or the character floats as soon as the camera pulls back.
Duplicating a wall layer, rotating it 90 degrees at the floor line and dropping opacity to about 28% fakes a convincing floor reflection.
Clicking a motion path keyframe with the pen tool converts it to a Bezier curve, which is often the whole fix for a hand move that feels jerky.
Scaling the background up while the camera pushes in creates a dolly zoom inside a 2.5D parallax shot.
A transition can carry the story: cutting from the heroic horse painting to the realistic mule painting tells the viewer the first image was propaganda.
Takeaway
Depth comes from preparation, not effects.
WHAT TO LEARN
Painted parallax looks real when each layer is cleanly separated, individually textured and rigged at true joints before a single camera keyframe is set.
03Upscale and extend the paintings
No AI model is reliably best at extending paintings, so generate several versions and composite the strongest sky, ground and edges by hand.
Price matters at scale: a model that costs 14 credits per image can still be worth it when the cheaper one keeps failing.
05AI separation with Higgsfield
AI layer decomposition is fast but caps at 2K, regenerates the image and picks its own splits, which makes it a poor fit for 4K work.
06Manual separation in Photoshop
Manual separation is one repeatable loop: select, duplicate, mask, Remove the object, merge, then clone-stamp the soft patches.
Split distant and near background elements into separate layers whenever there's real distance between them, or the parallax collapses.
07Texturing in Photoshop
A texture laid over the whole finished comp stays still while the scene moves, which instantly reads as a screen filter instead of paint.
Clip each texture to each separated layer and merge it in, so the grain travels with the image the way a real canvas would.
Texture only after separation, because textured images make generative fill and object removal noticeably worse.
08Character: separate the body parts
Cut a character into upper arm, forearm, wrist and hand, and paint back everything each piece used to hide, so rotation never reveals a hole.
Leave parts that won't move attached to the layer below; the horse keeps the rider's legs because the legs never animate.
09Organize and import into After Effects
An organizer script and a layered-PSD importer save hours, and live-linked PSDs mean fixes in Photoshop appear in After Effects instantly.
Rename and parent limb layers before rigging; skipping it is what forces extra helper layers later.
10Rig with DUIK Angela
DUIK Angela's auto-rig only behaves when every anchor point sits exactly on its real joint.
Undo after every test pose, because moving the controller and leaving it shifts the rig's starting position.
11Animate the eyes and head
Painted figures have no neck, so a head that wiggles needs a painted neck, a gradient shape layer or a duplicated, masked head piece behind it.
Blinks can be built by sandwiching eyelid layers between an add-masked and a subtract-masked head, then looped with time remapping.
Low wiggle frequencies like 0.3 to 0.5 read as natural breathing sway; higher values look nervous.
12Animate the hands
If a hand move feels jerky despite eased keyframes, convert the motion path point to a Bezier curve with the pen tool.
A front coat layer with a soft drop shadow sells a hand slipping inside clothing far better than masking alone.
133D camera: the basics
Animate nulls, not the camera, and chain them so each move begins before the last ends.
Wide lenses exaggerate parallax and long lenses flatten it; pick focal length to match how much depth you want.
Render slow, steady camera moves at 60fps, since 24fps produces visible judder on painterly parallax.
14Place layers in 3D space
Scaling an orbit null pushes a layer back in depth without changing how it looks from the camera.
Ground a standing figure with a floor plane rotated -90 degrees, corner-pinned to match the painting's perspective.
Fake reflections and shadows with rotated, low-opacity duplicates and shadow-only drop shadows masked to the floor.
15Camera animation
When nulls are chained, editing the first keyframe of an upstream null shifts every move after it, so only touch the keyframe you mean to.
16The sword transition
Transitions work best when they carry meaning; the sword flight and the camera both peak in speed at the same moment.
A track-matted duplicate whose opacity fades is what reveals the next scene through an object's shape.
A CC Lens warp keyed during the push-in makes it feel like the camera is flying through the object into the new world.
17Napoleon on the horse and the mule
Breeze adds wind to hair and cloth from a simple mask, with a negative mask expansion on the original layer hiding the seam.
A background that scales up while the camera pushes in gives a dolly zoom inside an otherwise flat parallax shot.
Cutting from the heroic horse painting to the realistic mule with a fire luma matte tells the propaganda story without a word of narration.
Glossary
Terms worth knowing.
Parallax
The effect where near objects appear to move faster than distant ones as the camera moves. In 2.5D animation it's faked by placing flat image layers at different depths in 3D space.
Layer decomposition
An AI feature that splits a flat image into separate layers automatically. It saves time but regenerates the image and gives you little control over where the splits fall.
Generative Fill / Remove
Photoshop's AI tools that fill in the area behind an object you've cut out, so the layer underneath isn't left with a hole when the foreground moves.
Clipping mask
A Photoshop setup where a layer only shows inside the pixels of the layer directly below it. Used here so a texture only appears on one cut-out object.
DUIK Angela
A free character rigging toolkit for After Effects. Its auto-rig turns a chain of parented limb layers into a limb you can pose with one controller.
Anchor point
The pivot a layer rotates and scales around in After Effects. For a limb or head, it has to sit on the real joint or the rotation looks wrong.
Null object
An invisible After Effects layer used only as a handle. Parenting the camera to a null and animating the null gives smoother, more controllable camera moves.
Orbit null
A null After Effects creates from a camera so the camera can be moved and orbited as one unit. Used here to push layers back in depth without changing how they look.
One-node camera
An After Effects camera with only position and rotation, the simplest to animate freely. A two-node camera adds a point of interest it always looks at.
Wiggle expression
An After Effects expression, written as wiggle(frequency, amount), that adds random organic motion to a property, such as a slow head sway.
Track matte / Luma matte
A way to use one layer's shape or brightness to decide where another layer is visible. Here it reveals the next scene through a sword shape and through a fire overlay.
Breeze
A paid After Effects plugin that adds wind motion to masked areas of a layer, used here on horse hair and cloaks.
CC RepeTile
A built-in After Effects effect that tiles a layer's edges outward, used to hide hard edges when a layer is pushed back or scaled.
Dolly zoom
A camera trick where the camera moves toward a subject while the background scales the other way, making space appear to stretch.
“The AI gives you the options, but it's on you to decide which one looks right.”
Clear stance on AI as a tool, not the artist.→ TikTok hook↗ Tweet quote
28:05
“A real texture is embedded in the image. When the image moves, the texture moves with it.”
One-line fix to a mistake most viewers are making.→ IG reel cold open↗ Tweet quote
2:20:10
“For smooth camera movement use nulls. Don't animate the camera directly.”
Short, absolute rule with an instant payoff.→ TikTok hook↗ Tweet quote
2:20:31
“Slow movement at 24fps creates motion jitter. At 60, it's smooth. Yes, it's more render time, but it's worth it.”
Answers a common comment and states a trade-off plainly.→ newsletter pull-quote↗ Tweet quote
3:54:52
“He wasn't on a horse. He was on a mule.”
A history twist told through a visual transition.→ IG reel cold open↗ Tweet quote
1:30:51
“Because all our layers are named and parented properly, auto-rig just works.”
Preparation beats tools, said in one line.→ newsletter pull-quote↗ Tweet quote
The Script
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
17px
metaphoranalogystory
In this video, we're going to recreate this animation I did. I'll walk you through my whole process from Photoshop to final render in After Effects, step by step. Here's a quick recap of what we're gonna cover.
The story and the research behind the paintings, upscaling and extending the images, separating images into layers in two ways, one with AI and one manually, texturing, rigging and animating the character in After Effects, the 3D camera, and placing the layers in 3D space, the small transition, then the next two scenes.
They play on a horse and on a mule. Timestamps are in the description. So head over to whatever section you need.
Project files for this tutorial and the original are up on computer for a small fee. Link is in the description. And if you wanna follow along, I have made a starter file with all the assets already in place.
This fire transition I made, these birds flying through overlay, expressions, scripts, presets, all of it that's gonna be used in this project is available there. All linked below.
After Napoleon becomes First Consul, he heads for his first major battle, the Battle of Marengo. To get there, he leads the reserve army across the Great St. Bernard Pass in the Alps in May 1800 to suppress the Austrians in Italy.
For that, we have source three famous paintings. The first, Napoleon as First Consul by Benoist. The second, The Crossing of the Alps by Jacques -Louis David.
That one was a propaganda piece. In reality, he crossed on a mule and 50 years later, Delaroche painted the realistic version. that's the third all three are in public domain i'll link everyone in the description now we have got the images but we can't use them yet they are not in the aspect ratio we need for the animation 3840 by 2160 so we upscale and extend carry painting before anything else this is the one step where we can't be lazy The AI model I use for most of my work is nano vanilla pro.
It gives me the most cinematic and accurate results for my work. But this time when I used it to upscale and extend the images, the results were just not that good. A lot of AI models can do this and honestly none is the best.
Which one works depends on the individual cases. So you have to test a few. So I used GPT image 2.
It does tends to be on the expensive side though. A single image is 14 credits compared to nano vanilla pro which is just 4. It did a pretty good job for the first and third painting.
Here's how it actually goes. You give it the image, ask it to extend the canvas and the result, not that good. So you run it again and again until you enter with the first six versions.
This is where Cover Human Instincts comes in. The AI gives you the options, but it's on you and your card to decide which one looks right. And often no single version is perfect.
So you take the sky from one, the ground from another and combine them into the final image. One more thing at this stage, for the first concept painting, I generated a top -down version of the ground, just by adding the extended painting as a reference and writing a simple prompt, that answers these questions that I get asked in the comments a lot.
We'll come back to this later in the video.
Now, for the parallax, we need to turn each image into layers, foreground, mid -ground and background. For this project, I did everything manually in Photoshop. But there are two ways to do this now.
So I'll show you both. The first is AI. Higgs field has a layer decomposition tool that takes a flat image and automatically splits it into layers.
And you get separated layers just as easily. Go to edit and upload your media.
We'll select this first council image. Then let it prepare your project.
Now click on the layer decomposition tool. One thing to note, it only works up to 2K, not 4K. So select 2K here.
Then choose how many layers you want. We'll go with 8 for this one. And 8 layers at 2K resolutions cost 24 credits.
Here are the results and you can see it has separated out all the layers.
you can download them all at once as a zip file extract it and here we have all our layers canva has a similar tool But this method comes with its drawbacks. 1.
You don't control where the separations happen. 2. The whole image gets regenerated in the process.
So, you are bound to lose some quality or notice small variations from the original. 3. It's only available up to 2K resolution.
So, you can't really use it for 4K high -res work. And the cost. On Hexfield, one 8 -layer decomposition is 24 credits.
That too in 2K resolution. On a longer project with dozens of images, that adds up fast. Now let's import the image.
Now we'll make a selection. Click on select subject in the contextual taskbar. As you can see, it has selected the character, but it has also picked up the table, which we don't want for now.
So we'll go to the quick selection tool here. You can click the minus button up here or just hold alt and drag over the table to deselect it.
While doing that it has also deselected the legs, but we'll add those back later So left click on the mouse to add to the selection and hold alt to remove the parts you don't need The more precise your selection the better the cutout will be Take your time here and keep the selection clean.
I am removing the rest of the table here and now adding the character's legs back in.
This looks good, but there's still a bit of a table left here. So I'll grab a quick selection tool. You can also press W to select it and remove that part.
Then click the mask button in the contextual taskbar. You can see there's some transparency on the character here. So I'll select that part again.
Now that mask looks good. Next, duplicate the layer.
Rename the underneath as character.
On the top layer, right click the layer mask and choose add mask to selection and then disable the mask. Now we'll hit remove. Leaving the table on its own without the connector.
This takes a little time.
It has done a pretty solid job, not perfect but good enough.
now select both layers right click and select merge layers we'll rename it later now with this layer selected click select subject again this time it picks up the table and then click on the mask button now we have the table mask duplicate the layer again with ctrl plus j Rename this one as table.
On the layer above, repeat the same process. Add mask to selection, disable the layer mask and hit remove to remove the table. Will repeat the same steps for every layer.
As you can see, it has removed the table and done a pretty good job. Now, merge layers.
But look here on the feathers, this area looks blurry.
So we'll grab the clone, stamp tool, hold alt to sample from feathers and paint over the blurry part.
That looks good. Now select the layer and click select subject again. And this time it has selected the objects in front of the wall, which is exactly what we needed.
Duplicate the layer and rename it.
Then again, add mask to selection, disable the layer mask and click remove.
Select subject is doing really good job this time. When I made the original version, I had to do a lot of manual tweaking on the selection, but it's handling it much better now.
It has removed the object in front of the wall. Again, not perfect removal, but it's usable.
So let's merge the layers and again you can see this part of the wall is missing. So we'll use the clone stamp tool, hold alt, sample from the left side of the wall and paint it over the right side.
Try to match it closely so it look consistent.
This edge still looks soft, so I'm sampling from the left side of the wall again to keep it consistent.
Now if we try select subject again, it says it couldn't find a subject. So this time we'll make the selection manually Select the whole area in front including the ground then click mask in contextual taskbar duplicate the layer again and rename it wall on the top layer right click add mask to selection disable the layer mask and hit remove This one didn't come out well.
Let's try it again This one looks good We can use it so let's go with this and merge the layers And let's make a proper selection we could have used this whole thing as a background but you can clearly see there's a lot of distance between the castle in the back and this tree and the structure so we'll separate these two we'll name it bg2 and the last one will be bg1 that gives a scene real sense of depth and it will work much better for the parallax you can see there's some transparency here so we'll select that part again that looks good now let's add this part that got left out and this looks pretty good now duplicate the layer and name it pg2 then on the layer again add mask to selection disable the layer mask and hit remove This looks good.
The whole background won't be visible in the panel version anyway. Only part of the castle will show. We just needed enough to fill the area behind what we removed.
So it's usable.
Merge the layers again and rename it BG1.
Now let's arrange everything properly. BG1 goes at the bottom, then BG2 and everything else stack up to the character on top.
The layers looks good.
Now let's save the file. Go to savers and save it as a psd with layers.
We'll be making adjustments to all of these layers next. So let's save another copy as a backup. That way we don't have to redo all of this if something goes wrong.
We'll try things out in this copy without touching the original file. Now let's import the other images.
From this image we will only use the sky as the background so let's remove the napoleon and the host.
Now this works perfectly as the background. Merge the layers and then import the next image which we will use for foreground.
Click done from this one will take the foreground For character will use the original painting instead since it's in a better quality So let's select the horse here And hit remove As you can see, the sky in this one isn't that good, which is why we are not using it as a background.
Now, let's select the soldiers, Nebulance Army.
You won't be able to select this part very precisely especially around the details like the guns and that's fine. Just get it as plain as you reasonably can.
I like the small bits over here too and remove the extra parts. This should work. Click mask and it looks good.
Rename it foreground.
for the character we will use the original painting we downloaded select subject and it made a clean selection and you can see it's in much higher resolution and better quality now let's import the third image Here we'll take the character from this image itself. So click on select subject It also selected the guy in the background.
So let's remove him from selection with a quick selection tool while holding alt Then add the rest of the character including the stick the person near the mural is holding the selection looks good so lets mask it and rename it then we follow the same steps as before duplicate the layer add mask to selection disable the mask and remove the character That gave a weird result, so let's try again This one's decent, but let's try one more time This looks good Now, merge the layers.
Next, let's select the foreground with quick selection tool.
same steps again duplicate the layer rename the layer underneath as foreground then on the top layer add mask to selection disable the mask and click remove The result isn't great here.
delete the layer mask we just created and deselect the extra part we had selected click mask again this looks decent and usable rename it army then add mask to selection disable the layer mask and hit remove on the above layer this looks decent it just needs a little fixing so we'll use the clone stamp tool again Now let's arrange the layers properly.
Alright, this looks good. Now let's save these three files.
Most textured animations we see do it the same way. Finish the animation. Put a texture layer on the top.
Set it to overlay. Done. And you can see the texture just sits there.
The camera moves and the texture stays perfectly still on the top. Like it's on a screen. That works if it's supposed to be a lens start.
But it's not. That's not how the texture works. A real texture is embedded in the image.
When the image moves, the texture moves with it. So the texture goes on each layer in Photoshop. The only reason we don't do it before the separation is that it would give us not so good results in the generator fill.
Now let's do the texturing.
Just copy it from here and paste it into our file.
Arrange it properly so the brush strokes cover the painting and keep it on top.
Now go to the adjustment add the levels bring this sliders towards the peak here. So the press strokes become more prominent You can see the difference now Let's add the curves and increase the contrast a bit more.
Now hold alt and left click to create a clipping mask for this adjustment. That way they only affect the texture layer and nothing else.
Then set the blend mode to overlay and adjust the texture opacity now let's see if the texture is there it's subtle we don't want it to be too prominent let's group this and rename the group paint Now let's bring in the next texture Unlock it Paste it here and arrange it properly Now we'll copy the adjustment from this pen group and paste them outside the group And again hold alt and left click to clip mask them.
Then set the textures blend mode and opacity.
we also need one more adjustment here let's select hue and saturation hold alt and left click to clip this one too bring the saturation down Make sure human saturation layer sits below the other two.
Now it looks good. Group this and rename the group wood.
Now let's import the third texture, the halftone.
copy and paste it here we don't need to do much with this one just resize and arrange it properly Then change the blend mode and opacity and that's it for this texture.
Now I don't want the character to be textured I want the character clean with the texture only on the background and other things So we have moved the character above all the textures Let's group everything and name it texture One more thing at this stage, we changed the opacity and blend mode of each texture inside its group.
But you can also do that on the group itself. Here I have set the texture to 100 % opacity and now I am changing the blend mode and opacity on the group instead. So if you are texturing a single image, you can control everything through the groups and keep it all organized.
That's not how we'll do it here though. We need to clip each texture to each individual layer and grouping doesn't work for that So now let's duplicate the texture group and keep it as a backup now we'll merge the texture and Its adjustment into one layer since we need single image for clipping mask rename it paint and bring it out of the group Then select the opacity and blend mode again here.
Delete the empty group and do the same for the next one.
The halftones stays the same as it was. Just set it up properly.
This looks good. Just adjust the opacity.
Now we'll clip this texture to the layers. I have moved the panel over to the left so you get a clearer view.
Move the halftone texture down and duplicate it four times, one for each layer.
Place halftone copy above each layer and create a clipping mask on it. We only need to create clipping mask on the halftone layer because any layer we bring in between the halftone and the layer below gets clipped automatically.
So, let's duplicate the other two textures and drag them in between the halftone and painting layer.
You can see they are automatically clipped too.
Since we'll be animating these layers in After Effects, having all these texture layers won't be efficient so we'll merge them into each layer.
duplicate the texture and keep them in our backup texture folder as we'll use them to texture the other images now let's open the other files Copy the textures from this folder and paste it into all these files.
As you can see, position isn't right here, so let's fix that. Make sure you have the right layer selected while moving so you don't accidentally move the layer above it.
Now let's do the same on this image.
We won't be texturing these layers again, let's just copy the layers and paste them into this file. We can keep the extra layers as backup, but I'll delete them.
The character sits on the top of the texture.
Now the same process again hold alt and left click to create the clipping mask on the halftone.
Then bring these layers in between so they are clipped automatically.
Then merge and rename them.
Same with this one.
This one's already done so I'll save and close this file. Now let's import the ground we created for the first image and texture it here.
Let's export it as a png and name it floor And that's it all our images are textured Now let's separate the character first save this file and rename the layers to character click ok we don't need these other layers so delete them now we already have a selection here but look at the hands there's some transparency so let's add the mask to selection and disable layer mask That gives us the full image back.
Then delete the mask and we'll redo that area.
Select this part.
And this part. That adds this part back in. Then click mask again.
Now there's no transparency or see -through areas. It looks good.
Now we can use this character Duplicate the layer and keep it as a backup We'll work on this one now, let's remove the unwanted parts here near the hand with the razor tool You can see there's still some see -through area.
So select that part again and mask it.
Now this looks good. So let's apply the mask. Next, we'll separate the Nebula's right hand.
Duplicate the layer.
Then use polygonal lasso tool to make a selection around the right hand.
Now we have the hand here, so let's mask it. On the layer below, we'll use the eraser tool to remove the hand.
Increase the brush size and erase it.
Now the hand is removed from the body and we have it on its own layer.
Let's add the body layer. Select the hand layer and use the clone stamp tool to paint the hidden parts so the hand looks coherent. I am sampling from the hand itself and painting over this area.
This looks good.
Let's place the hand below the body layer and rename it hand 1.
Next, we'll separate this area from the hand. Again, with the polygonal lasso tool, make a selection around the wrist.
duplicate the layer on the top layer hit delete to remove the selection on the layer underneath first apply the layer mask Then hit mask in the contextual taskbar. So now the wrist is on this layer and gets deleted from the other one.
Now we will sample from the hand again and paint over the area we deleted.
Now place it below the wrist.
Our right hand is separated. Next the head.
Again, with the polygonal lesser tool, make a selection around the head.
duplicate the layer and rename it head on the layer below hit delete and on the head layer click mask that way the head is removed from the layer below and mask on the head layer now let's use the clone stamp tool again make sure you have the right layer selected sample from the body and paint over the neck area so it looks proper when the head moves in the animation.
Then clean up the edges around the neck.
It looks good. Now let's separate the left hand. Select the polygonal lesser tool and make a selection around the arm.
without the hand itself because we'll take the hand from another image.
Render the layer below as body and hit delete on it.
Then on the arm layer, click mask.
Now select the body layer and with the clone stamp tool, paint over the deleted area.
Increase the brush size sample from here and paint over this area.
I'm trying to remove this halo. It's showing up because the brush hardness is too low.
I'll also paint over this hand since we'll be using the hand from another image.
increase the brush hardness, sample from the nearby area and paint over it.
For this gold belt sample from the belt itself and print it in.
Sample from the right side of the waist Then from the left side and try to match them This looks good, just this part isn't working that well, so we'll paint over it.
This looks like a mess, but it's going to be behind the character's arm anyway, so it'll work. Still, let's fix this area a bit more. We'll try to get a clean body line here and remove the extra parts.
Here I am trying to give the character a proper waist and remove this aloe we got.
Now with the clone stamp tool make a selection around the sword and paint in that area above so it looks like a complete sword.
Now apply the mask and separate the arm into forearm and upper arm.
Since we'll be animating this, it's better to have them separate.
Duplicate the layer We'll paint over this part so it looks coherent.
Now we have our arm split.
This looks good. Now let's do the same for the right arm. Apply the layer mask.
And split it into forearm and upper arm.
Leave the selection on one layer and create a mask on other Then on the layer below use the clone stamp tool to extend the area Now let's do this image. First save this one as character too.
Now we will select the Napoleon's body hair. The way he is sitting on the horse.
Make sure the whole Napoleon is selected.
Now click mask obviously wherever the horse overlaps him that part will be missing. Don't worry. We'll paint that in later Now duplicate the layer and name it character And click mask on the layer below we'll select the horse And also include the Napoleon's body here since the Napoleon layer will sit above the horse layer that part sort itself out and we don't have to paint out his lower body and fill in the horse behind it.
If the legs were going to move we would have to do that but we won't be moving the legs so keeping them there works fine.
Now create a mask on the horse and we have it here With the eraser tool remove this below the horse's neck With the bronze stamp tool, paint out the hand on the horse's mouth. We will be animating the hair so we want it clean.
This looks good. So this one will work.
Now select the character layer first we'll separate the head from the body Select the head Duplicate the layer and rename it head.
On the head layer, apply the layer mask first then create a mask.
On the layer below, we'll erase this part of the head with the eraser tool.
Hide the head layers so you can see what you are erasing.
After that, we'll paint in the cape behind the head.
First, let's clean up the edges of the hand, so they are sharp and clear.
Then with the clone stamp tool, paint in that area.
While doing this, other parts will show up in the sample area too. So make sure you just select the other area to sample and paint.
I'll clean up the edges a bit.
Now apply the layer mask on the character layer.
With the clone stamp tool. We'll paint the cape into the area where the horse was.
sample and paint this area properly again this looks messy and not properly painted but it will sit behind the horse so no one will say we just need something there because when we animate the area behind can't be blank near the hand take your time and keep the edges clean The character is all separated as you can see it looks good now we will separate this hand too since it will be moving select it duplicate the layer first then create the mask I think we missed a finger here, so let's select that too.
On the layer below, hit delete to remove the selection. Then paint over the related area sampling from the nearby cape.
Turn on all the layers and it looks good Let's just fix a few edges around here Arrange all layers properly and that's it This is the original file of the animation we are going to create.
I have all the files here, so let's organize them. I am using this script to organize them all. As you can see, it has created all the folders and sorted everything.
Now, let's make the two folders.
One I'll call original and put all these folders in it.
The other is where I'll import all the assets for this tutorial.
Now let's import the psd files. All of them at once. For that, I made this script, which lets you import all the files together with layers that's retained, something you can't normally do in After Effects, so you would need to import them one by one.
And I have already included these in the starter file. Link below.
It gives us the composition plus the folders with all the layers.
Now if we open the composition you can see all the layers are exactly the same in the same order as in Photoshop And if you make any changes in Photoshop, they'll show up here, too Now let's bring in the two images one with Napoleon's eyes closed and one with his hand out In this image, you can see the eyes are closed. We'll use these eyelids while animating his eyes.
So let's take Napoleon's right eye, which is on our left side of the image and create a mask around it.
center the anchor point and now adjust the scale so it matches our character's eye this looks good now duplicate the layer to lift the mask and mask his left eye.
Mask it precisely And again center the anchor point and position it so it sits right over the eye. Adjust the scale until it looks right.
The right eye lines up and the left one too. Now let's rename the layers properly. You can see on screen how I am naming them.
Hand 1 is Napoleon's right hand and hand 2 is his left hand.
Now for the left hand as I said earlier we will be using the hand from another image Let me line it up with our character Lower the opacity and scale it to match the character so the hand doesn't look too big or too small, then we'll cut it out.
As you can see, I am moving the anchor point towards the character's legs and scaling it to roughly match the character's size.
Now increase the opacity and mask this hand.
Move the anchor point to the wrist since that's where we'll rotate it from when animating.
Fix its rotation around 4 degrees and this looks good.
Now place it with all the other left hand layers.
This wrist we have here is from the right hand. So let's place it below near all the right hand layers Now rename this as left wrist Also make sure to parent everything in this order the wrist to the forearm and forearm to the upper arm Don't skip the renaming and parenting as it makes rigging much easier later.
Now let's rig the character. We'll start with the right hand. Select all the right hand layers and center their anchor points.
Since they all came from the same image in the Photoshop, their anchor points are already in the center of composition.
So, select each layer one by one and move its anchor point to the joints.
Now, we have placed the anchor points on all three layers. Let's open Duke Gangela.
You can rig all kinds of things here, humanoids, birds and much more. It's really a solid plugin. But we'll use auto -rig.
If we hadn't separated and set up hands layers like this with all parented properly, we would have to use the other ways instead. So let's click auto -rig. And you can see it's done a solid job with just one controller without all the extra layers you usually get with Duke Gangela.
But look here at the elbow joint. It has got a breaking point. Let's try moving the anchor point lower and see if that fixes it.
That doesn't fix it. The anchor point needs to be at the right place, which is elbow joint. So let's move it there.
So far it's working really well, but there's still a breaking point. You can see it when I stretch it this far. But there's a easy fix to this.
Just extend this upper arm part with the close temp tool in the photoshop and it will be done.
We can control the wrist rotation separately here, while it still can be controlled with Duke controller.
So, it's really a clean setup without pulling up the extra layers.
One more thing, whenever you move the controller to test the rig, press Ctrl -Z to put it back. Otherwise, you'll mess up the controller's starting position.
Now here, I selected the left -hand layers and clicked Auto Rig without placing the anchor points first. And it's all over the place.
So, let's press ctrl z, select them again and place the anchor points properly at the joints.
The wrist anchor point was already in right spot. Now select them and click auto -rig. As I said before, because all our layers are named and parented properly, auto -rig just works.
And you can see, it moves properly.
Let's do the eyes animation now. Bring the head layer above the eyelids and create a mask around the eye.
Now duplicate the layer so you can see the mask. And on the layer below, set the mask mode to none and create a mask for the other eye.
This looks good. Now delete the extra one.
On this layer, set the mask mode to add for both. Then duplicate the layer and set the mask mode to subtract.
Now bring the eyelids in between these two layers.
Now we have them here. You can see there are some see -through areas here. So let's adjust the eyelid mask so it covers them properly.
Same for the other app.
Turn on motion blur and now let's place a position keyframe at around 10 frames.
And in the start move the eyelids above.
And around at 40 frames copy the first keyframe and paste them.
Then press F9 for easy ease.
I'll just use this preset I have built in Flow.
The right eyes animation doesn't look right. So let's delete it and redo it.
Copy the position keyframe from the left and paste it here. You can see the right eye needs to go a bit lower than the left. So let's do that.
Copy the first keyframe and paste it here.
Now we'll make the graph same on both.
Set the influence of the first point to 60 in both input and output velocity.
And keep the graph same on both layers.
I think it's already 6060 here.
On the last keyframe set the boot to 66 and 66.
For the middle keyframe set the incoming influence to 84 and outgoing to 74.
And do the same on other layer.
Now the graph looks nice and the animation on both eyes looks just the way we want But you can see there are still some see -through areas it's an easy fix Select the top head layer, the one with the subtracted mask Increasing it makes the mask bigger So we'll use a negative value, minus 2 That does a pretty good job now there are no see -through areas but now you will notice the lower part of the eye is showing so let's extend the lower eyelid mask a little so it's hidden Same for the other eye.
Now it looks perfect. Don't extend it too much or you'll start seeing the background. Looks perfect.
Now copy the keyframes and paste them again at around 100 frames.
And again at around 200.
For the third blink, we want it to happen fast like two quick blinks.
After that, we won't be copying the keyframes. We'll just loop this animation from around 7 second mark. Let's watch it one more time.
For this double blink let's rearrange the keyframes a bit so it looks right. Just bring these keyframes forward a bit.
This is the eyes animation from the original version You can see I did a lot of complex stuff like animating the upper and lower eyelids separately and all that stuff We don't need all that here As you can see this already looks pretty good Now rename this layer as eyes.
Then select the eyes and eyelids layers and pre -compose them as eyes.
Now enable Time Remapping on it.
Delete the last keyframe and then alt click the stopwatch and paste this loop expression.
The number in the expression is the duration in seconds you want to loop.
Here we want the loop every 7 seconds.
So we will set the number to 7.
Now you will see, it loops after every 7 seconds.
I accidentally brought this layer control in front, so let's move it back.
Now we'll do the character animation and we'll be covering this much.
First, parent the eyes to the head.
Then move the head's anchor point to the neck, so it rotates properly from there.
But you can see there's a gap here when the head moves Obviously, there's no neck and we can't do it like this This is the original version comp And what I did to solve this is create a shape layer for the neck.
It acts as the character's neck and fills that space.
So let's do that here. Select the paint tool. Set the fill to linear gradient with no stroke.
And draw a shape for the neck.
Pick the colors from the skin near the head.
You can adjust the gradients with these sliders.
Or in the properties, you can also use the starting point value to adjust it.
Try to match roughly like mine.
Then bring it below the head layer and parent it to it. Then duplicate the body layer bring it above and create mask around the neck part.
Neck color isn't blending that well. So let's reselect it You can see this dark edge of the head layer here Let's create a mask around the head so we don't see this edge And set it to intersect.
And then, just do some adjustment until it looks right.
Now, it blends properly.
Now, let's add a wiggle expression to head's rotation.
wiggle 1 comma 16 and you can see it starts to rotate but it feels like it's moving too fast So let's use 0 .5 instead of 1. It rotates the same amount just a bit slower.
Now you can see that when the head moves left, this neck part doesn't look right.
Let me move this body mask and see if it helps.
That doesn't look right. Let's try fixing it by adjusting the neck layers.
But it still doesn't look right. It feels like a cutout.
Let's adjust this mask a bit and make it look like the original again.
So here's what we'll do. Duplicate this neck mask layer. Delete the mask.
Create mask on a white colored edge around the neck.
And bring it here to create a proper edge.
Then adjust the mask.
Center the anchor point on the mask. rotate it and extend the mask to get a clean edge.
Now it looks good.
Now let's clean up these edges a bit. You can see these red lines here.
This is just nitpicking. But now it looks good. Now we have the head and eyes animation looping throughout.
This is the original animation.
We just need to set the keyframes on the controllers and the animation works.
You can see how I did the animation here. That's what we will recreate.
In this animation, one hand moves up and back down, and the other comes to his waist and then goes inside his coat.
Let's do that here. As I said earlier, press Ctrl Z to return the controller to its original position. Cause resetting the position will reset it to the center.
Then set the keyframes.
Looking at the original it moves up and forward around 2 seconds. So we will do the same.
At around 6 seconds the hand comes down.
So copy this keyframe and paste it here.
Then at around 12 seconds slide the hand forward a bit.
The hand is resting on the table so it just slides to the right.
Then easy is the keyframes with flow.
The hand needs to start moving up before the camera starts moving which is why I brought that keyframe forward.
Now for the left hand, this is the final position when it goes inside the squad at around 12 seconds.
We'll rotate the wrist later. First, let's block out the hand's movement.
In my original animation, around 2 seconds the hand comes near his stomach and then around six seconds it comes towards his sword and waist.
At around 10 seconds, there is an in -between keyframe for the movement, where his hand goes inside his coat.
That's what we are doing here. Set the keyframe here.
Now easy is the aim and lets play the animation.
It takes a while to load the preview.
It's playing slowly because my playback frame rate is set to 30. While our composition is 60. So on your end, it might look a bit faster.
But here his vest doesn't look right. The body shape looks off.
So let's fix it with a mask to get a clean body shape.
Select the pendulum and create a mask like this.
Set it to subtract.
Now adjust the mask a bit.
This looks decent, but the body still looks a bit boxy and square Let's move this point here And the top one here too This looks decent Bring this in a bit more Adjust this part a bit Now it looks good.
Now let's animate the wrist. We'll rotate it and add keyframes around it for the left hand.
And at 12 seconds, add a keyframe.
At the start, rotate it and add a keyframe.
Hide the controllers for now.
And we will just add a few keyframes in between so it matches the hand movement.
In the original you can see how I did it.
Here it should feel like he is resting his hand on his waist or it's near his fort. Right now it's too low. So let's bring the hand up.
On the wrist we are seeing the cut so lets just bring it back.
Now we have brought the hand up and its on the waist.
Adjust the wrist rotation and also set an in -between keyframe for the wrist rotation.
And easy is it.
Now the right wrist.
Set a keyframe at its original position. First, move the controller back a bit so we can see the first frame.
Now, set a keyframe at this point.
And then Z1 where it points towards the text.
Then set a keyframe at the end and adjust the hand controller a bit.
Move this layer back and ease all the keyframes and bring it forward again.
You can see, this animation isn't feeling smooth. It should feel like a continuous motion. So let's play around with the graphs.
Adjust the graph like this. But it still doesn't feel smooth. In the original, it really feels like he is putting his hand inside smoothly.
Here it doesn't. So what's the problem?
It's just a one simple fix.
With the pen tool, click on the keyframe point on the path. You can see it created a proper cloud brazier path.
Now it feels smooth.
Now we will adjust the wrist keyframes.
Delete this keyframe and just bring this one back.
Adjust the rotation a bit so we don't see the cutout.
And adjust the graph here so it rotates properly as it goes inside the code.
And bring it forward a bit.
Extend the mask on the wrist and bring it back.
Now it rotates properly with the animation.
Now duplicate the body layer Delete this mask And mask the coat here that will sit in front of his hand let's rename it as body in front and place it above the hand you can see the hand below now So let's adjust the wrist rotation a bit and then adjust the position of the hand with the controller and adjust until it looks right.
Now it's perfect.
And then add a drop shadow to it so it casts a shadow on the hand.
Adjust the direction and place it like this.
Add the similar drop shadow to the wrist too.
Adjust the control again so it looks right. Again this depends on the personal preference.
Here this elbow joint is not looking right and looks pointy.
So let's fix this.
Select the layer and create a mask.
And set the mask to subtract.
and adjust it a little to create a proper shape it looks good Now even if we stretch it far it doesn't break After effects gives you two kinds of cameras a one note camera is like a camera on your shoulder It has position and rotation.
And that's it. So it's the simplest to animate freely. A two node camera also has a point of interest.
A target itself is looking at. Move the camera and it keeps pointing at that spot. For parallax I mostly use one node.
Focal length works like a rear lens. A short focal length say 24mm is a wide lens. So small camera moves gives you a lot of parallax.
A focal length say 85 or 100 lightens the scene. And you get less separation for the same move. For smooth camera movement use nulls.
Don't animate the camera directly. Parent the camera to a null and animate the null. Then stack the null above it and parent the first to the second.
Now make sure the animation on the second null starts before the first ends. Use the graphs to adjust the speed. The overlapping nulls gives you that smooth and continuous camera movement.
The reason for comments like this on my animation is 60fps. Here's why. Most of my camera follows story spacing.
Slow movement at 24fps creates motion jitter, and it looks unconnected to the eye. At 60, it's smooth. Yes, it's more in the time, but it's worth it.
Now, let's create a new composition. Same as before, 4K 60fps. And duration of around 30 to 40 seconds works for us.
Now we have our comp here. As you know when we imported the files you got the folders with all the layers and compositions. So everything is already sorted.
If you want to import layers from these you can. You can also select them from the compositions we got here.
Let's open the first console composition. You can already see it's properly layered. Copy the layers in sequence and make sure to copy from top to bottom.
Then paste them here.
For the character, we'll import the character composition, the one where we did the animation.
You can see our character is here and the animation is playing.
Let's go to the start.
Close the other account. Right click, go to the new and create a new camera. For now, we'll use the standard 35mm camera.
No need to change anything else.
Click ok and we have our camera Now click the toggle switches button and turn on motion blur and 3d layers for all Switch to two views and set the second to top view And also turn off the transparency grid.
Now select the camera. Right click and choose create orbit null.
You will see it has created a null with the camera parented to it. Now unparent the camera and parent the null to the camera instead.
Then select the null's position and set the Z value to 0.
You will see the null's position jumps to the camera's position. Let me show you again.
Set the Z value to 0 and it moves to the camera's position.
Now let's take our background parent it to the null. And now scale the null. As you engage the scale you will see in the top view that it moves back.
While in the camera view on the left nothing changes.
So I'll scale it and it looks exactly same from the camera's perspective.
You can see it has moved back. It's at 5600. Let's make it 5000.
Now parent the bg2 to the null and scale it. You can see on the right that it moves back too.
Then select the wall and front of wall layers and parent them to the null. And scale them to the position that looks right.
Now unparent the front of wall layer and scale it a little. So there is some distance between the wall and the front of wall layers.
Now unparent all and parent the camera to the null instead.
As I change the null's position, you can see we are getting proper parallax.
But when I move back like this, you can see the character's legs are floating. There's no floor here, which is exactly why we made that floor image. Now let's reset the camera.
I'll press control Z to get it back to its original position.
Just delete the null and now import the floor.
Turn on motion blur and 3D and set the X rotation to minus 90.
Then bring it down just enough so it sits under the character's feet.
This looks good. You can also switch to second view to check it.
Let me go to the custom view. If you rotate around here, you can see all the layers are properly separated and our floor is there too. Now let's focus on the floor, the reflections and the shadow.
First, let's create an orbit null for the camera. Now we have our floor here. So we'll try to match to the floor we see in the painting, its perspective and the way it looks.
Let's switch to two views and set this one to left view.
Now bring the floor back so it roughly sits at the wall space.
Match it roughly and bring it down a bit so it sits right on that edge.
Now let's add a corner fill effect to the floor.
With the floor selected, make sure you are selecting the pins themselves and position the upper right one like this.
Stretch it out so we roughly get the floor's perspective.
The upper left and right are done. Now go to the lower right and stretch it like this.
Now the lower left. I accidentally selected the layer instead, so press Ctrl Z and make sure you click on the pin before stretching it.
This looks a bit too high above the wall's edge, so let's bring it down a little.
This looks good. Now we will add a CC reptile effect to extend the left and right edges.
Turn on the unfold in tiling and increase it.
You will see there is a gap now. To fix that drag the reptile effect above the corner pane.
Then expand the right and left sides.
Now there are no gaps and it's extending properly.
Next select the wall there and duplicate it. On this copy we'll rotate it by 90 degrees.
But first move its anchor point to where the wall meets the floor.
Bring it like this Then set the X rotation to 90 Let's take it in the second view and see how it looks Let's scale it a bit on Y axis only.
Now duplicate the front of wall layer 2 and do the same. Go with anchor point to where the wall meets the floor and set the X rotation to 90.
You can try other blend modes here like overlay or screen, but they usually don't give good results So we'll keep it on normal and just lower the opacity Now it looks like the wall is reflecting on the ground This looks decent. Let's raise the opacity a bit.
28 looks good.
Now let's rename these layers adding reflection to their names so everything stays organized.
Next add a drop shadow to the table.
Let me show you why we are doing this. Here's the original painting. Let me move this panel over here.
You can see there's a shadow under the table and we are roughly matching it.
So let's increase the distance a bit.
I am adjusting the opacity, distance and softness until it looks right.
These values look good. Now duplicate the table layer.
Move its anchor point down to the base again and we will create a reflection from it.
Rename it table reflection.
Delete the drop shadow from it.
And rotate it by 90 degrees on x.
We'll need corner pin on this one too. To get the corners right, let's set them like this.
Lower the opacity to about 25 % Now move it below the table layer You can see part of the table is getting cut off at the ground. So let's bring it up a bit This much looks good. Don't bring it up too far Next, we will create a mask for the table shadow.
For that duplicate the layer Rename the layer table shadow and put it below.
Remove the drop shadow from the table layer. Then set the drop shadow on this layer to shadow only. Then create a mask.
Now you can see the table isn't casting shadow upwards or onto the objects behind it. The shadow only falls on the ground.
I'll adjust the mask a bit until the shadow sits right.
Now it looks good change the values a bit and see what looks best 6 in the softness is already perfect. Let's adjust the mask until it looks right Save the file now, let's add the birds The birds will sit in the background.
So let's go towards the background. You don't have to match my camera position or camera movement here. We'll cover camera animation in the next chapter.
But look at this area. It doesn't look good.
This is the result of bad generative fill.
So let's head back to the original file in Photoshop and fix it with clone stamp tool I'm sampling from the nearby areas and printing this out That's not working in the view So let's try taking this window and painting it here That's not working either. So let's paint this area from here instead That's still not right so let's extend this area it's fine to select small bits and extend them gradually Now it's starting to look like a proper age Let's straighten this edge.
It's looking bent.
Now it's perfect. Just fix this a bit more for a clean edge.
You won't see this in the close up. It is the only in white shot. So make sure it looks right from the distance.
You don't have to worry about the lower part. It won't be visible much. And you can see in after effects it's already updated.
Now let's import the birds.
Here in overlay folder we have the bird sequence which I rendered out from blender. I created flying animation there and exported it as a png sequence.
Make sure to check this, create composition and import.
And here's the animation playing. It's moving slowly because my playback frame rate is set to 30.
Now bring the birds comp into our comp.
Turn on motion blur and 3D for the layer.
Then press Ctrl Shift and parent it to BG1. That way it takes BG1's position and sits properly in 3D space.
Now add tint to darken them.
Then add inbuilt flip preset And set scale width to minus hundred Here scale the birds layer down a bit Now let's animate them flying out Set a keyframe here And this is the keyframe where they fly out of frame. Then drag this keyframe until the speed of them flying out looks right to you.
Now let's work on the camera animation. First change the camera's focal length to 44. Something around 44 to 45 felt right to me.
Now bring the camera back a little so we get a proper view.
Now you can see this is the right position but we need to extend the floor down a bit.
same for the reflection layers then copy and paste that to others now let's create an orbit null and set its color to yellow also push the cache before we start so we have enough memory And set the playback frame rate to 60.
Let's set the z position to 3800. Something like this.
And adjust a bit.
Let's zoom in a bit, so we have birds in focus.
And also I noticed for the birds, I accidentally selected the stress layers, so deselect that. Now they look perfect. We want the birds in focus for the first shot.
So that's what we are setting up here.
Set the keyframe on the null, then move to the second position.
Not this much, just so the character is visible like this.
There isn't much to explain here, so just follow along with what I'm doing on screen.
I chose this easing curve because we want the camera to start fast at the front and then slow down as it arrives here.
Now I feel his hand is a bit too low to be pointing at the text box. So let's go into the character comp and bring it up a bit.
and also adjust the wrist rotation. That way it points correctly at the box we'll add here.
Now it looks perfect.
The next camera move should start before the end of this one.
So let's duplicate the null, put the playhead where you want it and press the left square bracket.
And the layer moves to the playhead.
Now we have our layer, delete its keyframes and parent the null below to this one.
Now we want a close -up of our character. So increase the Z value and zoom in on him Position it like this with the character centered And again we'll ease it with my normal easies preset It needs to start a bit sooner so bring it forward.
Now, this move starts just before the previous one ends. We'll do the same again for the next one.
now select and delete the keyframes and here set a keyframe and parent the below layer to this one now bring it towards the right side and one important thing don't select the first keyframe while making these changes all the nulls below are parented up the chain So if you move the first keyframe on the null above it will mess up the animation below it.
Here we want a graph like this. So the camera moves right exactly as his hand goes inside his coat.
That matches nicely We duplicate the null again bring it here And delete its keyframes Now add the first keyframe and go to the second position And here you can see I have mistakenly selected keyframes highlighted in blue and I am making a change. Watch the problem it causes.
You can see it has shifted the whole position of our previous animation.
So press Ctrl Z and make sure that first keyframe isn't selected.
Now make the change on this keyframe and move it back.
The camera comes down, moves here, goes across and back. That's perfect. Now duplicate it again and bring it here.
Delete the keyframes. This will be the last move. Set the keyframe at start position and parent the layer below to this one.
Here we want the camera to pull back so our sword comes into frame. Let's bring it back Now we are starting to see the floor and also the edges on the top Let's bring it down like this We don't wanna go below the ground and that looks perfect.
Let's easy easy and that's it. This already looks good.
Now let's just load the preview and see the animation.
The camera animation looks good. Now let's work on the sword animation. So let's import this word png.
This is the word we generated and it's textured the same way as everything else. One thing here, make sure to uncheck create composition while importing since we just want the file itself.
Bring it into comp and turn on 3d for it.
Now I am pressing ctrl and page up which centers the layer in view. That's the shortcut I have set up for myself.
Go to edit keyboard shortcuts search for center layer in view and see what it's mapped to on your system.
Once it's centered we can work with it properly Now set the keyframes on position scale and rotation Let's switch to two views so we can see where it sits in there and push it back like this This much should be fine now look at the left view Bring it in left like this This much looks good and adjust a little Now scale it down to 40 and bring it below the frame so it comes from underneath Add a rotation keyframe Then at this point it comes up into frame up to here Bring the scale value to this keyframe.
So the scale stays same until this position And then change the rotation to something like this around minus 38 to minus 40 Then bring it forward it should come towards the camera like this Select the pen tool and we'll click on a point on the motion part to turn it into curved bezier part.
And then add a keyframe here to lift the path up so the sword arcs over the table instead of going through it.
bring it up a bit more and do the same on this one.
now ease them bring this rotation keyframe like this so it turns as it travels and here we are adjusting the path handles of the back point like this and bring this keyframe forward a bit.
Now select all position keyframes and we will work in the graphic editor.
Now what we'll do is bring these keyframes up like this.
But first, center them on this keyframe, click continuous and pull the handle up just a little. Do the same on this one, continuous and pull up like this.
So the graph reads as just two clean pics. Right now we are not getting a clean pic here. So what we'll do is delete this keyframe and work on this too first.
We'll add it back later.
Let's try to get a bit of a preview.
Adjust the graph a bit like this Now with the pen tool add a point here on the path and bring it up Go back to the graph and you can see it now sits on the path, which is what we wanted So the motion carries through properly and smoothly Something got messed up here. So ctrl Z and bring it up Now this looks good bring it up near the character's head Now select the rotation and let's fix it Here you can see it's cutting through the wall So first let's adjust the Z position on the first keyframe There's no fixed set up or formula for this, it's mostly figuring out as you go.
Also adjust the rotation of it.
Now let's push this point behind the wall Let's watch the rough preview Now select this keyframe and move it like this Then select the rotation keyframe and adjust the graph of it So it rotates and arrives like this and move the last position slightly to the left.
So the sword is in center.
Now let's fix the path again which changed when I brought that point back and adjusted it.
And here it's still crossing the table and the character. So we'll lift it up and adjust the scale keyframe.
The animation looks solid here. I'll bring in the second scene from here.
Now just extend the out point and at this point we'll have the second scene started.
From here what we want is our camera will fly through this ward and reveal our next scene.
This is the composition for that scene. So let's bring it in as it is. We will use to build the transition first and later it will animate the character and whole camera moves inside that composition itself.
Turn on 3D for it. Then center layer in view and scale it down.
Now we need a camera push in for the transition. So duplicate the null and parent the null below to this one.
Now let's push the camera in. By that I mean increasing the Z position so it moves into the shot.
For now, leave it like this.
But looking at it, I feel the sword is sitting too far forward and too big in front of the camera. So let's turn off the second scene for now. We'll come back to it and fix the sword first.
I have closed that composition and now I'm adjusting the sword's last position so you can see it properly in the view.
Now for the mask, go on frame ahead and duplicate this word layer.
Set the mask mode to subtract. We will duplicate the layer below with the mask on it and it will act as a reveal. So now let's bring in the comp back.
Turn on 3D for it and scale it down.
And center it so it sits properly in the composition.
For the transition, what we want is that when the camera zooms in, the second composition should Fill the whole frame. So position it properly now.
So it ends up full screen as the camera comes in.
Take a moment and get the second scene composition properly centered.
Create a new null object for the sword.
Sword is going to rotate and that rotation is what reveals their scene. So take the second scene and set it as a track matte to the sword layer, the one with the mask. Then duplicate that mask layer because that's what creates the reveal.
Here's why. If we just lower the opacity on the matte layer or the second composition layers, there's nothing left to reveal through. The cutout part simply won't be visible.
So instead we duplicate it and animate the opacity on this copy.
As the opacity comes down it reveals our scene.
Now add a rotation keyframe on the null. Turn on 3D for it as well and parent all the sword layers to this null.
So bring it forward to where it's at zero.
And at this point we want it to rotate a full 90 degrees.
And we would also want the sword to be centered. So we will also adjust the position.
Now select this and bring it back.
We want the sword to travel behind and out of frame. So move the camera like this.
Here it was sitting behind the character. So let's bring it forward. Just enough that it's in front of him.
Then scale it down.
You can see it travels through like this which is exactly the motion we wanted. Now turn on this layer and you will see the reveal. There it is revealing the second scene.
It just needs scaling up so it fills the whole composition. Like this.
It's sitting a bit high so bring it down like this so it fills the space properly Now this looks good But we are seeing the edges, bring that last keyframe on the reveal opacity forward a bit, so it reveals faster.
And now, let's expand the edges of the second cob, which is the reptile again.
But we are still getting a hard edge on the right side here and increasing the tiling wasn't fixing it. So the problem is actually the mask. So let's fix the mask instead.
I am copying the mask I fixed and pasting it onto all three layers. So the mask stays identical across them.
Now as it moves in like this, we want a bit of edge warp here. So it feels like the camera is flying through the sword and entering our scene. That's what I did in the original.
So let's add the CC lens effect on this.
If we set the size to 180, you can see it creates that warp look.
So let's set convergence to 0 and size to 180 and we'll put keyframes on it.
Now set the convergence to 100 and add a keyframe.
And from this point the camera start pushing back in the next scene.
So let's put marker here to mark where the camera move should start.
And then the camera push back from there.
Now let's work on the second scene. Turn on 3D for these two layers. Now let's bring in the character composition for the second scene.
Nephilim on the horse.
Turn on 3D for that too. And you can see it comes in much bigger size than we need. So let's scale it down.
And match it to its original position in the painting.
Try to match it to its proper position and resize it You can see at 40 it's going out of frame So let's import the extended image we made bring it into the composition And use it to check roughly where the character should see And how big it should be. Then delete it again and scale our character down to around 35.
That's perfect. Bring it down into position like this.
Now let's create a new camera.
For this comp we will use 35mm camera. Again nothing different.
Then create an orbit null And parent the null to the camera the same way as before Then zero its position parent the background to it and scale the null It's the exact same process as the first scene And you can see The background separating out in depth Now unparent the layer to the null and reset the null's position And parent the camera to the null also change its color to yellow Next let's go into our character composition because we'll animate the character in there for that we need a bit more room So let's change the width of the composition This much 8968 is fine because we'll be creating mask and we need space to see what we are doing Turn on motion blur for all the layers Now move the anchor points of the hand and head to their proper pivot points like this And rename this layer to horse since that's what it is.
Then on the character layer duplicate it and bring the copy above.
Here I am masking out just the hand because that's the part that needs to sit above the horse layer.
This looks good.
Now let's animate and add wind to the horse's hairs.
You can see the horse layer here. Let's isolate the horse and mask his tail.
Like this create the mask now go to the breeze plugin pick the breeze 2 preset With the wind direction set to left and hit apply Play it back and you can see the movement in detail It already feels like wind moving through the horses hair Let's set the wave width to 350 and wave speed to 8 Then on the layer below use mask expansion the same way we did earlier in the chapters Set it to negative value.
So it hides the edge of the cutout we just made Here's what Breeze is actually doing. When you create a mask on the layer and hit apply, it duplicates that layer with the mask applied and puts all the wind movement on that duplicate.
Your original layer stays intact with just the mask set to subtract. So a negative mask expansion is what adds the seam and you can put multiple masks on the original layer.
so here i have made a second mask and set the mask expansion to minus 21 on this one also in the original layer now i am doing the same for the upper hairs you can see i accidentally started the second mask here so ctrl z And make sure you keep working on the original mask.
Now the mask is made. Hit apply. On base again.
It creates a third layer with that mask applied. Set the wave width to 350 and wave speed to 8. On this one and on the previous one too.
That looks perfect.
And again set the mask expansion to minus 21 for the third mask as well.
Let's isolate the character. We'll add wind to the knippling's red cloak.
Create a mask for it the same way. Also turn on the head and hand layer so you would know where to draw the mask from.
Set the wave width to 350 and wave speed to 9 and again set the mask expansion to minus 21 Now same for the part of cloak behind his back Since that would obviously catch the wind too, so let's mask that mask it like this heat apply on breeze set mask expansion to minus 21 and you can see the cut appearing as it moves this always going to be some we just want it as low as possible So change the wave height to minus 4 and you can see it's still there but much less visible.
Let's try 13. No, that's worse. So let's keep it near minus 4 and set the wave width to 350 and wave speed to 8.
Let's see how it looks.
But let's try 4 instead of minus 4.
Yeah, that's better but bring it down a little And this is perfect, so I'll keep it at 2 and we wave it to 350 and we'll speed to it Let's play it through That looks good, but I'll bring it up slightly to 3 Now adjust the mask a little And that works.
Now select the head and add wiggle on the rotation. Same as before. Wiggle 0 .5 comma 16.
By the way, the negative sign doesn't change anything. It behaves exactly the same.
And on the hand, 1 comma 16.
But it feels far too fast So let's try wiggle 0 .5 comma 15 That looks decent but the hand is still moving a bit too much So let's go with so let's go with 0 .2 comma 8 instead Yeah, that looks right for the hand.
Now duplicate the character layer and bring the copy above.
Delete its masks.
And create a new mask around his neck collar. So that part stays in front of his head.
That way the caller sits in front of the character head rather than behind him.
Let's bring the hand down slightly so you don't see that red patch.
And here you can see we have got cut out again. This character doesn't have a neck either. So let's head over to Photoshop and paint one in.
Apply the layer mask first, then use the clone stamp tool, sample from his face and paint in the neck area.
Lets also extend the hand a bit so it holds up.
And clean up the edge of the hand so it doesn't look bumpy.
This looks good and as we save it you can see it updates straight away in After Effects Now the mask I made on the collar isn't great rather than patching it let's just delete it and redo it properly this time so we don't run into same problem so let's make a precise mask This part is fine as it is.
Now change the head and hands wiggle a little to 0 .3 ,8.
That works for the amount of movement we want here. After seeing the preview, I decided this value works best for both.
Now select all character layers and pre -compose them. Since we want one overall movement on the horse.
Name it animated.
And deselect this open new composition.
Now we'll use these again to add a movement. Create a mask like this, leaving out his legs or at least the lower part of them. Mask it like this.
Use this slow -motion preset with the wind direction set to above and two -way raise selected and click apply.
You can see it added a movement.
Change the view height to 15. Now there's a motion and it looks right.
That's perfect.
Now let's do the camera animation just before the marker will set a keyframe and start the move there So, let's push in on the character We want the character centered here, so let's move it like this And check it in our final comp It feels like it's sitting a bit too far left So let's reset the position for now now increase this reposition and bring it down like this so we have our character setting here then also increase the scale of the background that gives us that dolly zoom effect first set a keyframe at the original value bring it back then add a keyframe on the increased background scale.
After that around 10 seconds in bring the camera back and add a keyframe.
Now ease the keyframes on both.
And the camera move looks great. We are getting a dolly zoom here.
Bring these keyframes forward a bit so the camera moves faster.
Now increase the out point a bit so we have room for the third scene which comes later.
From this point on we don't need to see the layers below or this word So hold alt and click on left square bracket and trim them to here But don't select the camera will still need it for what comes next Now I am loading the previews to see how the whole thing is looking Let's pause the preview.
Here you can see the camera start zooming out a little too late So let's select these keyframes and adjust them like this in the graph Now let's preview it again And it looks perfect.
Adjust the graph so the sword arrives just as the camera reaches its fastest point pulling out.
Now, let's fix the sword's rotation a bit, so it rotates properly as it reaches the wall and moves in front.
I am testing things out here, so just follow the final graph shape.
And this looks good to me.
Now lets add an adjustment layer and put a tint on it to darken the background as the sword comes into frame.
Set the keyframes on the 10 to 15 then add a fast box blur and set it to 2 Now let's add CC light sweep on the sword.
Now you can see the light swimming across as the sword comes in.
Frame the adjustment layer and the null we created for the sword.
and extend this sword layer the one our second scene is track matted to extend it so it stays until the end and from here we'll start on our third scene napoleon on the mule Now duplicate the null and bring it back like this.
Delete the old keyframe and add a new keyframe.
And parent the null below to this one.
What we are doing with this camera move is bringing the Napoleon on the horse into a centered view like this.
And in the middle of this camera move. We'll transition into the third scene. Napoleon on the mule with the fire transition.
So let's input the fire transition and the snow overlay which we'll need for the third scene.
Both are included in the starter file Bring the both fire transitions into the composition and Set the blend mode of the fire transition 2 to add. We'll use below one as a track matte for all the layers As you know we have all the third scenes layers separated into their own composition and sorted into folders.
So let's open that composition, copy all the layers again from top to bottom and paste them here. Then trim them with alt and left square bracket. For now, hide the fire transition and turn on motion blur and 3D for all the layers we just imported.
Now copy background's position from the second scene bg and paste it onto the bg2. We'll also copy the scale but we don't want the keyframe so just paste the final value on the bg2.
And you can see it's already sitting in a position. Now for the foreground, the character and the rest they are already moving as you can see.
But nebulon on the horse is noticeably bigger than the nebulon on the mule.
So let's parent all the layers to be the character layer. And scale it. So both nebulons end up roughly the same size.
takes us from the second scene into the third scene. It reads properly and carries the message. He wasn't on a horse.
He was on a mule.
Now select all the layers of the third scene. Set them to luma matte to the fire transition. And turn on invert so it reveals properly.
Then select the foreground, background and the character layers of the second scene. And set them to luma matte as well. But for these, don't click invert.
As you can see, the second scene vanish as the third one appears.
Now let's adjust the inel's final position a bit. Our character sits on the left in this third scene. So we want the framing to suit that.
You will also see I am adjusting the background as I go. Which I will keep doing throughout. Just nudging the background so it doesn't look off.
And the transition is working nicely. We are getting those fan elements coming through from that transition.
Also, make sure to push the media cache so you get a smoother playback on the preview.
Now let's import the snowfall we need for the third scene. Bring it like this.
And scale it. Set it to screen.
Set it to luma matte on fire transition 2 so it only shows up once the third scene starts Now let's add another orbit null And a new keyframe For this move we want to bring the mule into focus, so let's get it centered You can see background for this scene is sitting part 2 back. Which just looks weird in this ground area.
So bring it forward a bit like this.
I am taking my time to getting it right. I actually intended to build this scene using projection mapping at first.
And even ran a few tests.
So I ultimately decided to stick with this technique instead.
Now I'm getting a proper preview at the quarter resolution. But from what I can see, this looks decent enough. And you can see the background is still a bit too far back.
So let's bring it forward.
Now it's become too big. So let's scale it down If I bring it down, even with the reptile, the upper part doesn't look good, which isn't what we want.
And we'll need a mask here because the mule's back legs are supposed to be hidden behind the foreground layer.
So create a mask like this so the mules back legs stay hidden and the front legs stay visible.
Adjust the mask like this. This looks good.
You can see the leg isn't quite right here. The same as in the painting itself. Which is exactly why we are doing this.
Go to the end keyframe and change its position like this. Our end goal is to focus on the mule. So let's frame it like this.
The mule is visible and we are not showing too much ground.
That looks perfect.
Now let's animate the head. We'll add a head rotation the same way we did for the other characters.
So let's mask it here.
Obviously, this character doesn't have a neck either.
So you can head into Photoshop and paint one in or you can create a shape layer instead.
Now move the anchor point of the head to its proper pivot point for the rotation.
And add a wiggle.
Same 0 .3 comma 8.
I think it's too slow. So change the value from 8 to 16. Now that looks right.
Duplicate the layers. Rename this one the character in front.
And duplicate once more and bring that below as our body.
Remove the rotation expression from both the layers and delete the mask on the character in front layer.
And on the same layer mask the neck collar area which will sit in front of our head layer.
Now on the body layer instead of deleting the mask set it to subtract.
First, let's work on the mask. Adjust it like this so it covers the whole area and doesn't cut out.
Now, as I said, instead of using a shape layer for the neck, here's a better way.
Hide the shape layer, duplicate the head layer and bring the copy down a bit like this.
Delete its mask, then create a new mask on it for the neck area and shape it properly.
On the top head layer, move its mask like this so we don't see a harsh edge and it blends properly.
Then adjust the mask on the neck collar a little so it forms a proper neck shape. There's nothing special happening here. Just a small adjustment to get the shape right.
Also add a feather on the masks so the two blends together properly. That's it.
Now let's animate the clothes that's flying out with Breeze. The same process we followed for the second scene.
Select the layer, create a mask.
Set the preset to breeze 2 with the wind direction set to right instead of left and hit apply.
Then make a mask on the left side too and follow the same process.
Set the mask expansion to negative value on the original layer.
Make sure to turn on the second scene layers back on before moving forward.
Now let's change the wiggle values on the head. Set it to 0 .5 ,8 on both the head layers. I found those values work much better.
Now adjust the neck mask a bit more to get a clean shape like this.
It reads as part of the face now. That's what we wanted.
Here you can see the mask feather is pulling it a bit far. So let's bring it back below like this so it doesn't cause any problems.
Adjust the neck part once more and that's it.
Now let's fix a few things I spotted in the preview you can see this edge showing on the sword That's from the mask So we do the same as before set the mask expansion to negative value on the layer above the one with the mask set to subtract I also notice the light sweep is a bit too fast So I am adding a keyframe in the middle and bringing it like this to adjust the speed That looks right to me.
You can copy the position values you see on the screen for this I also noticed we wanted this fire transition to happen at the camera's peak speed as it's moving forward So I have set the camera's peak to land here and that's where the fire transition happens We never added an X rotation on the sword. We need that too because it makes the animation look a lot better.
So let's add X rotation on the sword.
0 at first and the last keyframe and minus 44 in the middle.
Then adjust the graph for that rotation so it comes up properly and the movement reads smoothly I also noticed the birds are bit too big. They are meant to be in the distance.
So they should look smaller. Let's scale them down.
And we still want them in focus at the start. So let's zoom in a bit more on the first keyframe. You only select the first one.
If you select the second keyframe as well on this null, it'll mess up all the animation keyframes that follows.
Now let's do the text box. There's nothing complicated here. It's just a shape layer with a gradient and textures on it, plus a few effects.
Let me turn it off so you can see all the textures that's on it It's a rectangle with roughen edges and the drop shadow nothing special So, let me show you how it's made Click on the radial gradient option and draw a rectangle like this If you open up the rectangle's properties You will see there's a stroke And below that a gradient fill.
Go into the gradient fill and you can see the gradient handles. Moving this pointer changes the gradient on the layer. So let's set our pointer like this.
Then for the gradient colors, I'll just copy from the original. You can see the hex value on the screen.
Copy this color and paste it.
I'll paste this same color here. But we want a slightly darker version.
This much looks right. And it matches the original.
Add a stroke on it.
Change the stroke width to something like 29 Then copy the color from the original and paste it in Then add roughen edges as you can see in the corners, that's the kind of fill we want Keep it on roughen edges and adjust the values Follow along on the screen as I make the changes.
Yeah, this weathered look on the borders is what I wanted.
Now add a drop shadow and set it like this. And our shape is ready.
Then the textures. Instead of going back into the photoshop to texture it, I did the same process here.
And used displacement. layer also it doesn't make a huge difference but i tried so it's in here for the text i used helvetica originally let's change it to clash display instead so we'll center it and use this font for the project Now, I'll copy the composition from the original scene and paste it here.
That way, it'll be at the same position on this one too.
Then again the second scene. I'll change its content. Copy it from the first original scene and just paste it here.
It's not complicated. If you have watched the whole video, you already know how to place a composition in a 3D space and position it properly. So I don't need to walk you through it step by step.
That's the full breakdown. If you want the project file, it's on Gumroad. Link in the description.
If you want more of this, please let me know in the comments. And if you make something with this, tag me at TheMohithArt on IG or X. I would love to take a look.
The Hook
The bait, then the rug-pull.
The finished piece opens on a sweep past birds over a painted Roman skyline and lands on Napoleon pointing at a map. Then the video rewinds to zero and rebuilds every second of it, one selection, one keyframe and one null at a time, for four hours.
Frameworks
Named ideas worth stealing.
06:50list
The manual layer-separation loop
Select subject (fix the selection with quick select)
Mask it and duplicate the layer
On the top copy: add mask to selection, disable the mask
Run Remove to fill the hole behind the object
Merge, clone-stamp the soft spots, repeat for the next object
One repeatable Photoshop routine peels a painting into depth layers while filling in what's behind each object, so nothing shows a hole when the camera moves.
Steal forany 2.5D parallax shot built from a still image
28:05concept
Texture per layer, not on top
A texture overlay on the final comp stays still while the scene moves. Clipping the texture into each separated layer makes the grain travel with the paint, the way a real canvas would.
Steal forany textured or painterly motion piece
1:29:24list
Rig-ready limb setup
Separate upper arm, forearm, wrist and hand
Rename every layer clearly
Parent wrist to forearm to upper arm
Move each anchor point onto the real joint
Run DUIK auto-rig
Prepared this way, DUIK Angela's auto-rig builds a clean one-controller limb with no extra helper layers.
Steal forrigging any cut-out character in After Effects
2:20:10model
Stacked overlapping nulls
Parent the camera to a null
Animate the null, never the camera
Add a second null and parent the first to it
Start each new move before the previous one ends
Shape speed in the graph editor
Each camera move lives on its own null, and the overlap between them is what makes the whole move feel continuous rather than stop-start.
Steal forany After Effects camera move with more than one beat
2:23:03model
Orbit null depth placement
Create an orbit null, zero its Z so it sits at the camera, parent a layer to it and scale the null. The layer moves back in depth while looking identical from the camera, giving true parallax with no reframing.
Steal forplacing flat images in 3D space quickly
2:19:41concept
Focal length and parallax
Wide lenses (around 24mm) turn small moves into big parallax; long lenses (85 to 100mm) flatten the scene. This project lands at 44mm.
Steal forchoosing camera settings for 2.5D animation
CTA Breakdown
How they asked for the click.
VERBAL ASK
4:18:26product
“That's the full breakdown. If you want the project file, it's on Gumroad. Link in the description. If you want more of this, please let me know in the comments.”
Soft and brief. The starter file is pitched in the first minute, then the close just points back to Gumroad and asks viewers to tag their own versions.
Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.
Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
A step-by-step walkthrough of training Claude to write, design, and animate a Vox-style documentary end to end, using three custom skills and Higgsfield to generate the images and animation.
Five prompting habits that turn a default, generic AI promo video into one that actually matches a brand's own look, from finding reference videos to reviewing drafts like a designer.
A 40-minute raw screen session where the only human work is a few typed prompts, and Claude Code does the cut, grade, graphics and sound inside DaVinci Resolve, After Effects and Blender.
A professional thumbnail designer rebuilds a trilogy's visual identity, generative-fills a hair cutout, and FaceTimes the creator for real-time feedback before the video goes live.
08:04
10:31
18:16
21:07
22:40
25:53
29:49
37:11
40:25
42:24
45:29
48:37
57:01
58:27
62:16
67:46
67:59
71:32
76:20
79:20
84:06
88:57
95:25
98:39
102:34
103:33
108:21
115:33
116:33
119:50
125:54
130:23
132:43
137:28
147:10
150:24
153:38
156:52
159:28
166:34
168:38
173:03
175:31
179:31
185:59
189:13
192:27
195:41
198:55
202:25
206:45
208:37
211:51
216:27
220:47
224:48
228:02
231:16
234:30
237:44
240:58
245:10
251:40
253:54
17:15
05:39
11:58
27:27
40:45
14:24