Layers and components

#Add a layer

Press + Add Layer under the Layers list. The Add a layer chooser opens with five tiles:

• Media adds an empty picture or video layer that covers the frame, ready for you to load something into it (see Media).
• Gradient adds a two-colour gradient in your brand colours, at 90% scale (see Gradients).
• Vector adds a vector layer with one glyph, for logos, icons, symbols and text (see Vectors).
• Particles adds a Vector that is already a still hex lattice of particles. Add Motion to it to make them move (see Particles).
• Points adds a layer of moving points that other layers follow. It never shows in the ad (see Points).

A layer's kind is fixed once it is added. To swap a gradient for a photo, delete it and add a Media layer.

The stack

The Layers list is in render order. The top of the list is the back of the picture and the bottom is the front, so a new layer lands at the bottom of the list, in front of everything else. Each row has:

• a drag handle to reorder it,
• an eye to hide or show it,
• a lock, which freezes its values so nothing can change them by accident,
• Duplicate, which puts a copy just behind the original,
• Bring to top, which renders the layer above the text and everything else,
• Delete. The Background has no delete: every scene keeps one.

A linked layer's type icon names the scenes it travels to (see Link), and its row shows a warning sign when a particle chain may jump at the cut; hover either for the details, or tap it on an iPad. The one link icon on a row is the pose link toggle among its buttons. Click the selected row to fold or open its settings; the components under it fold and open from their own headers.

Layers are named "Media 1", "Gradient 2", "Vector 3" and so on. Double-click a name to rename it; Enter saves and Escape cancels. A scene holds up to 12 layers as well as its Background.

Select: Click or Layer

The Select switch beside the layer count sets what a click on the canvas does:

• Click (the default) selects the object you click, so you can pick any layer straight from the picture.
• Layer locks selection to the selected layer. Clicks and drags on the canvas only move and transform that layer, even where another object sits on top of it, and nothing else can be picked there: the pointer shows a no-entry cursor over everything else. To work on another layer, click its row in Layers (or press Alt+Up or Alt+Down).

Wildfield remembers your choice in this browser. Select applies in Canvas mode; Layout mode picks its text boxes and CTA the same way it always has.

Shortcuts

• Cmd/Ctrl+C and Cmd/Ctrl+V copy and paste the whole active layer. Add Shift to copy and paste only its pose onto another layer.
• Alt+Up and Alt+Down select the previous or next layer.
• Cmd/Ctrl+Z undoes and Cmd/Ctrl+Shift+Z redoes, including adds and deletes.
• Z turns the zoom loupe on and off, and Esc closes it.

The keyboard button in the header lists every shortcut (on a narrow window or an iPad it is in the header's More menu).

Zoom loupe

The loupe is a magnifying glass over the ad, for checking small print, edges and detail at full quality. Press Z (not while typing in a field) or the magnifier button beside the ad's top-right corner to turn it on, and Z, the button or Esc to turn it off.

• A large glass circle follows the pointer over the ad at 3x. Scroll or pinch to zoom from 1.5x to 8x.
• The zoom is sharp, not an enlarged screenshot: text and vectors are drawn at the magnified size and images come from their full file. The rim bends the picture like real glass.
• While it is on, clicks, drags and scrolling on the ad are blocked, so you cannot move a layer by accident. Turn it off to edit again.
• The glass hides during playback and comes back when you stop.

Layers and scenes

Layers belong to a scene. When scenes are linked (with + Add Link), a layer you add appears on every scene of that linked family, with its own pose on each, which is how a layer travels between scenes (see Link).

#Add a component

Every selected layer ends with + Add Component. It opens the Add a component list: each option has a title and a one-line description, and it only lists what this layer can take and does not have yet.

What each layer takes

• Background: Image grade and Crop. A gradient Background also takes Shape and Gradient FX, like a Gradient layer.
• Media: Crop, Mask, 3D Transform, Image grade and Ease.
• Gradient: Shape, Crop, Mask, 3D Transform, Gradient FX and Ease.
• Vector: Count, Fill canvas, Fan out, Draw, 3D Transform, Mask (its Media fill: the shape shows a picture or video, see Media fill), Motion (with its Displace), Trail, Connect / Chart, Gradient Morph, Displacement and Ease.

On a Vector the list is grouped Layout, no motion, Movement and Timing. The particle modules all need Count, which turns the Vector into particles, so they are only offered once Count is there. Ease, the curve of the layer's own move between the poses of linked scenes, sits under Timing on every layer and is always last. Keyframes need no component: they are automatic (see Keyframes).

Removing a component

Close a component with its x and it is removed as if it was never added: its values go back to a fresh layer's. Closing Count also removes every particle module that needs it, and the toast says how many. It is all one undo step, and if you add the same component again in the same session, Wildfield restores your last settings for it.

Closing Ease puts the layer's move back to Auto. The layer keeps its poses, its keyframes and its links: they are automatic and belong to no card.

Waiting components

A component that cannot act right now shows a Waiting badge and says why on the card:

• the layer is still locked to the frame (press Make Free on the layer),
• the particles are off (turn them on in Count),
• a Mask's source layer was deleted (pick another),
• Crop on the Background only works while there is a Timeline Video.

Its settings are kept while it waits.

#Background

Every scene has a Background layer at the top of the list, behind everything. Its Source is either Media, a picture, or Gradient, two colours at an angle.

• A gradient background is shared by every scene on the aspect ratio you are editing, so changing it once recolours the whole ad in that format. Copy this gradient to the other aspect ratios sets the same gradient on the other formats. It has every control a Gradient layer has, Gradient FX included, so it can animate.
• A still image background belongs to the scene you are on. Use this image in all aspect ratios puts it on this scene's other formats, and the flip buttons mirror it.
• A Timeline Video plays under every scene and sits on the Background too (see Timeline Video).

The Background is always pinned to the frame. It takes Image grade and Crop (Crop only for a Timeline Video), plus Shape and Gradient FX when it is a gradient, and it never moves between scenes, so it has no keyframes and no Ease. Its Scale runs from 100% to 500%, down to 10% while a Timeline Video is on. Turn its eye off to see the scene with no background at all.

#Media

A Media layer shows one picture or one video. An empty slot has two buttons: Cloud loads from a connected Dropbox or Google Drive (connect one in Input first), and Local loads from this computer. You can also drop a file onto the slot or pick one in the Library.

• Images up to 25 MB and videos up to 200 MB.
• The same file is stored once. Upload a picture or video you already uploaded, under any file name, and Wildfield uses the copy it already has: nothing is sent again and the Library shows it once.
• A picture fills the frame but is never cropped on arrival. A landscape photo in 9:16 fills the frame by zooming in, and the parts the frame hides stay in the layer: drag it sideways to show them, or scale it down to see all of it. Every other aspect ratio fills its own frame the same way until you position it there. The picture is only cropped to the frame by Render Image and the export. A video fills the frame as before.
• Dropping on the canvas works too: an image makes a new Media layer, an SVG makes a new Vector where you dropped it, and a video fills the selected Media layer or, with no Media layer selected, becomes the Timeline Video.

A still image gets Render Image, Retouch Image and Extend Edges buttons under it (see Retouch). Render Image bakes the framing you see into a new picture at this format's size, so retouch and AI edits work on that crop. A video gets Trim (see Clip trim).

Revert to original

Wildfield keeps the picture you loaded into a layer (an upload, a drop, a cloud pick or a link) as its original. Render Image, Retouch, Extend Edges and AI edits never replace it. Once the layer shows something else, a round arrow appears at the top left of the picture in Media: Revert to original puts the original back, uncropped and unedited, at full-frame framing.

It works on the aspect ratio you are in and no other. So you can render a 9:16 crop and retouch it, switch to 16:9, revert, reposition the original for the wide frame and render a new crop there, while the 9:16 keeps its finished image. The rendered and edited versions stay in the layer's history (the arrows in Media) and in the Library, and Undo takes a revert back. A picture that was generated rather than loaded counts as the original once you first edit it.

#Position and fit

The layer card's position controls place, size and turn the layer:

• X and Y move it. Click a label to reset it to the centre.
• Scale sizes it: 10% to 1000% for Media and Gradient layers, and 2% to 4000% for Vectors. Click to reset it to 100%.
• Rotate turns it from -180° to 180°. Click to reset it to 0°.

You can also drag the layer on the canvas while a Drag to move badge shows. Reset puts every position value back, and the position lock freezes X, Y, Scale and Rotate so a stray drag cannot move it.

When the selected layer runs off the frame, a small map in the bottom-right corner of the canvas shows the frame and where the layer sits. Double-click the map to centre the layer in the frame: only X and Y change, its scale and rotation stay, and one Cmd/Ctrl+Z puts it back. It moves the layer in the current aspect ratio only (unless the layer is Relative), and a layer that follows a point keeps following it. A layer locked on its row does not move.

Absolute or Relative

• Absolute (the default) gives each aspect ratio its own position and scale, so the layer can sit top left in 16:9 and centred in 9:16.
• Relative stores the position as a distance from the centre, shared by every ratio. Use it for layers that should sit in the same place everywhere, such as a centred badge.

Locked to the frame

Some layers from older projects, and Live Frame captures, are locked to the frame. They show Locked to the frame. Make it Free to move, scale or animate it. with a Make Free button. Until you press it, Crop, 3D Transform and Ease wait, and the layer does not travel between linked scenes.

#Points

A Points layer holds moving points, called tracks, that other layers follow. It is a guide, not a picture: it never shows in the preview's playback, a render or an export, only in the editor. Select it and its open track is drawn on the frame.

Its card lists its tracks:

• Track picks which one you are editing, and Delete removes it. Layers that followed it stay where their own Position puts them.
• + Orbit adds a point that circles the frame (Radius, Turns, Centre X and Centre Y). + Sweep adds one that crosses from left to right at a Height.
• A face, hand or body tracked from a video is a track too (see Track a face, hand or body), and so is a point or an object you pick in a video (see Track a point or an object).
• Record adds a track from your pointer: a 3-2-1 count-in, then the scene plays once and the point follows your pointer. Rec records the open track again, Loop keeps every pass as a take (pick one under Takes), holding the mouse button while you record punches in over just that part, and Edit shows its points on the frame to drag, delete or add. Smooth evens it out.
• Name renames the track; the name is what the other layers' menus show.
• Used by lists the layers that follow the track.

Two things can follow a track:

• A particle layer's Displace magnet (Attract, Wake, Trail or Emit): pick the track under Path, in the Points group. See Particles.
• Any other layer's position: pick the track under Follow point, which appears on every layer once the scene has a Points layer. The layer then rides the point, and its own Position becomes its offset from the point, so X 0 and Y 0 sit right on it. A layer linked to another scene keeps its link's motion instead.

Tracks are timed by the scene, so a longer scene plays them more slowly, and placed by the frame, so they land in the same place in every format. Changing a track moves everything that follows it straight away; deleting the Points layer puts the particle magnets back on their default orbit.

#Track a face, hand or body

A Media layer with a video or a picture has a Track row with Face, Hand and Body. Press one and Wildfield finds it in every frame and adds it to the scene's Points layer (making one if there is none) as a track named Face 1, Hand 1 or Body 1. It runs on this computer: the footage never leaves it, and nothing is downloaded until you press Track the first time (the model is about 16 to 20 MB, kept for next time). The row shows how far it has got, and Cancel stops it without changing anything. The whole result is one undo step.

A tracked face, hand or body has several named points, and each one is listed on its own under Follow point and Path (for example Points 1 · Face 1 · Nose):

• Face: Nose, Left eye, Right eye, Mouth, Chin, Forehead and Head (the middle of the face).
• Hand: Index tip, Thumb tip, Middle tip, Wrist and Palm. With two hands in the picture it keeps the first one it finds.
• Body: Chest, Hips, Nose, shoulders, wrists and ankles.

Left and right are the person's own. A face too small for the face model, as in a wide shot, is found through the body first.

A tracked point stays on the footage: trim the clip, move or scale its layer, or look at another format and the point still sits on the face. Its card on the Points layer says which layer it came from and how many frames it was found in; frames where it was lost are bridged. Track again re-tracks the video (after a new trim, say) and everything that follows the track keeps following.

#Track a point or an object

A Media layer with a video also has Point and Object under Track. Both open the tracker over the preview: the clip, a slider to pick a frame, and Track and Cancel.

• Point follows one spot. Go to a frame where it shows well and click it. Pick a spot with some detail (a corner, a mark, a bit of pattern); a plain patch of colour gives it nothing to hold on to. It is quick, about 30 frames a second, and the first time downloads OpenCV (about 3 MB).
• Object follows a thing. Go to a frame where the whole thing shows and drag a box around it. Wildfield outlines it on every frame and gives five points: Centre (the middle of what shows), Top, Bottom, Left and Right (the middle of each edge of its box). It uses the SAM 2 model on this computer's graphics card, so it needs a desktop browser with WebGPU (Chrome or Edge on a Mac or PC). The first time downloads the model (95.6 MB, kept in this browser for next time), and it runs at about 2 to 3 frames a second: 10 seconds of video takes 2 to 3 minutes. Without WebGPU the tracker says so and offers Point instead.

Either way, tracking runs forward from your frame to the end of the clip and back to its start, drawing each frame as it goes, and Cancel (or Esc) stops it without changing anything. The footage never leaves the computer, and there is no charge. The result is one track on the scene's Points layer (Point 1 or Object 1), one undo step, and it stays on the footage through trims, moves and formats exactly like a face.

When the spot or the thing is hidden (behind something, or out of the frame) the track has a gap there rather than a wrong position, and the card says how long it was lost for. It is looked for again as it comes back, but a full hide can lose it. Re-seed on the card opens the tracker on the first lost frame, with the lost stretches marked in red under the slider: go to a frame where it shows, click it (or box it) again and press Track. Everything before that frame stays as it was, and tracking restarts from there. With no new mark, Track again tracks the whole clip again from your marks (after a trim, say).

While something is partly covered, an object's Centre is the middle of the part that shows, so it drifts; an edge that stays in view (Top, say) is steadier.

#Crop

Crop trims the edges of a picture, video or gradient. It has two sliders with two handles each: Top ↕ Bottom and Left ↔ Right, each from 0 to 100%. Click an axis label to reset it.

Crop starts locked so a stray drag does not change it: unlock it with the lock in the card header to edit. The header also has an eye to show or hide the crop handles on the canvas, a button to apply the crop to every pose, and Reset.

Apply Crop bakes the crop into a new, flattened image. Use it when you want the cropped picture as a clean asset to reuse, for example before AI retouching. Vectors have no Crop: size their artwork instead. On the Background, Crop applies to the Timeline Video only.

#3D Transform

3D Transform lets a flat layer lean into the scene: a phone screen tilted back, a card spinning toward the viewer, a poster receding into depth. It works on Media, Gradient and Vector layers, not the Background.

The controls

• Tilt X, Spin Y and Roll Z each run from -180° to 180°. Roll Z is the same value as the layer's 2D Rotate.
• Depth Z runs from -100 to 100, pushing the layer back or pulling it forward.
• Perspective runs from 20 to 400 (160 by default). Low values exaggerate the foreshortening, high values flatten it.
• The trackball is the quick way: drag it to tilt and spin, hold Shift to roll, and hold Alt for free movement. Grid and Snap 15° help with exact angles.

Click a label to zero that value, or press Reset to clear them all. The 3D values are part of the layer's pose, so on linked scenes the layer turns in 3D as it travels between them. Perspective is scaled to the frame, so a tilt looks the same in the preview and at every export size.

Worked example

A 9:16 app ad where the phone screen swings into view.

1. Add a Media layer with the app screenshot and set Scale to 70%.
2. Add 3D Transform and set Spin Y to 35° and Tilt X to 10°.
3. Press + Add Link under the scenes and, on the linked scene, set Spin Y back to 0° and Tilt X to 0°.
4. Play: the screen swings round to face the viewer as the scenes cut.

#Image grade

Image grade corrects and styles the colour of a picture or video layer, including the Background. It appears as Image on the layer card.

• Contrast and Saturation run from 0 to 200% (100% is unchanged). They apply evenly to the whole layer and never animate.
• Blur softens the layer from 0 to 20 px and can animate between poses.
• Brightness (0 to 200%, 100% unchanged), Hue (0 to 360°) and Grain (0 to 100) can animate between poses too.

Click any label to reset it. Gradients have no Image grade: their softening is Soft Edge in Shape. On a layer still locked to the frame, only Contrast, Saturation and Blur show.

Tips

• Drop Saturation to about 20% on a busy lifestyle photo and add a brand gradient over it, for a consistent campaign look across different photographers.
• A little Grain (10 to 20) takes the digital edge off AI images and stock photos.
• Animate Blur from 12 px to 0 across a linked scene for a focus pull on the product.