Animated Lower Thirds and Overlays for YouTube: Specs, Timing and Tools

Good lower thirds for YouTube follow five rules. Animate in fast (about a third to half a second). Hold still long enough to read the text twice. Animate out faster than you came in. Keep everything inside a 5% margin from each edge, clear of faces, captions and the player controls. Use contrast of at least 4.5:1. Build them in your editor, or with any tool that exports a transparent background, because a lower third has to sit on top of your footage.

Full-screen pieces work differently: intro title cards, chapter cards, quote cards and outro stings. They replace the picture for a few seconds, so any tool that outputs a normal video file can make them.

That split decides the tool. Oakgen Motion (our product) makes the full-frame kind: type-led cards and short clips in 16:9 or 9:16 that you cut into your timeline. It exports MP4 only, with no transparent background, so it cannot make a lower third that floats over your footage. For that, use your editor's presets.

The specs at a glance

The durations are our working ranges for talk-to-camera videos, not a platform rule; the end screen window is YouTube's. Sources for each number are in the sections below, checked on October 22, 2026.

Overlay or full frame: the question that picks your tool

A lower third is an overlay. Your footage plays underneath, and only the strap and its text are opaque. For that, the graphic must either be built inside your editor on a track above the footage, or come from a file that carries an alpha channel (transparency information).

Most exports don't carry one. Jitter's help page on exporting with a transparent background lists MOV (ProRes 4444), WebM and GIF as its transparent formats, and says transparent exports are "only available on our Max and Ultra plans" (per their site as of October 2026). MP4 is not on that list. Drop a normal MP4 on a track above your footage and it covers the whole picture.

So: if you need the face and the name on screen at the same time, the graphic has to be made in the editor or exported with alpha. If the graphic can take over the frame for a few seconds, any video file will do, and you gain more freedom with type, colour and motion.

Don't try to key the background out of a full-frame card to fake an overlay. Text edges fringe and soft motion turns muddy; rebuild the strap as a real overlay.

Safe areas for 16:9 YouTube videos

The most useful number comes from broadcast. The EBU's recommendation R 95, safe areas for 16:9 television production (version 1.1, June 2017) sets an action safe area of 3.5% and a graphics safe area of 5%, measured from the top, bottom and sides of the image. On a 1920 x 1080 frame, 5% is 96 pixels at each side and 54 pixels top and bottom.

YouTube doesn't crop like an old TV, but its interface draws over your video. Google says so for ads on its About video ad specs page: "Overlays, call-to-actions, and buttons may appear in different places depending on the format, campaign type, and screen," and Google offers transparent PNG safe-zone templates for horizontal, vertical and square video. Those templates were made for ads. We could not find an equivalent YouTube Help page for ordinary uploads, so treat the ad templates as a conservative guide, not a rule for your channel.

Then check the real thing. Upload an unlisted test with your lower thirds in place and look at it three ways:

1. Paused on a computer, mouse over the player. Anything the controls and progress bar cover is off-limits for text a viewer must read.
2. On a phone, tapped so the controls show, in portrait and full-screen landscape.
3. With captions on, at a large size. YouTube's caption settings let viewers change the font, colour, size, background and window of captions, so your viewers' captions may be bigger and boxier than yours. Your strap shouldn't collide with them.

Our default from that test: keep name straps inside the 5% margin and raise the bottom of the plate well above it, around 15% of the frame height from the bottom (about 160 pixels at 1080p). It's a starting point; move it after the test.

Safe areas for Shorts and other 9:16 video

Vertical video is harder because the interface covers more of the picture. YouTube publishes no safe-area numbers for Shorts that we could find. The only official placement guidance is in YouTube Help's Enhance your Shorts page: when you add text in the Shorts editor, visual guides "show you where elements like the like button, comments, and video description might appear for viewers," and white lines mark a "non-safe" area near the edges "where your content might be partially hidden on some devices."

What works without a number:

• If you add text in the YouTube app, follow those guides. If you edit elsewhere, use Google's vertical ad safe-zone template as a conservative overlay.
• Keep text, faces and the key part of each shot in the middle of the frame. The bottom and the right edge are where most of the interface sits.
• Re-lay out a 16:9 graphic for 9:16 rather than cropping it; the kinetic typography guide covers re-breaking lines for vertical frames.
• Watch the Short on a phone before publishing.

The same thinking applies to the cover frame; see our Instagram Reels cover guide for how covers get cropped in a grid.

Timing: in, hold, out

Every overlay has three phases, and the hold is the only one that matters to the viewer.

In. A slide, a wipe or a fade of a third to half a second (10 to 15 frames at 30 fps). Longer than that and the viewer is watching animation instead of reading. Ease out so the text arrives and settles rather than stopping dead.

Hold. Count from the moment the text stops moving, and hold for twice the read time, with a floor of about two and a half seconds. Reading-speed limits and the read-it-twice rule are covered in the readability rules for moving text; a strap also competes with a face and a voice, so lean long.

Out. Faster than the entrance, a quarter to a third of a second. Leave on a cut point or a pause in speech if you can; an overlay that leaves mid-sentence pulls the eye.

Frequency. Name a person the first time they speak, and again only after a long absence or a new segment. A strap on every cut becomes wallpaper.

A sample timed lower-third spec

Here is a spec sheet for an invented channel, "Tidewater Workshop", a furniture-restoration channel with one presenter. Copy the structure and replace the values with yours. Times are at 30 fps on a 1920 x 1080 frame; pixel sizes are font sizes and are our starting points, not standards.

The handle sits at the top and stays out of the last 20 seconds for a reason covered in the end screen section.

Type size, contrast and brand consistency

Size. A 16:9 video on a phone held upright is only a strip across the middle of the screen, so size for that, not your edit monitor. The pixel sizes above are where we'd start at 1080p; then watch the export on a phone at arm's length. If you squint at the role line, go bigger or cut words. Keep strap lines to around 30 characters.

Contrast. Aim for at least 4.5:1 (WCAG 1.4.3), and don't set text straight on footage: the background changes shot to shot. Put a solid plate behind it and test against the brightest shot it will appear over.

Brand consistency. Fix two typefaces, two or three colours, one entry direction, set positions and set durations, then save the strap as an editor template so a new one means typing a name.

What to avoid

Covering faces. A name strap across someone's chin is the most common mistake. Frame interviews with a strap in mind, leaving lower space on the side the strap comes in from.

Covering the subject. On tutorial channels, hands and the work are the content. Move the callout, or wait for a wide shot.

Fighting the YouTube interface. Controls, the progress bar and viewer captions all sit over your picture at times. The unlisted test above is the fix.

Stacking. A strap, burned-in captions and a callout at the same moment is too much text. If you burn in captions, raise or move the strap while they run.

Text in the end screen zone. YouTube places end screen elements over your last 5 to 20 seconds, so straps or handles there can end up underneath them.

Flashing. Glitch transitions and strobing reveals can trigger seizures in people with photosensitive epilepsy. WCAG's success criterion 2.3.1 says content should not contain "anything that flashes more than three times in any one second period" unless the flash is below the general and red flash thresholds. Keep flashy accents rare, small and away from saturated red.

YouTube end screen graphics

An end screen is a full-frame background with YouTube's buttons on top. You design the background in your edit; YouTube places the clickable elements. Per YouTube Help on end screens, as of October 2026:

• "End screens can be added to the last 5–20 seconds of a video."
• "Your video has to be at least 25 seconds long to have an end screen."
• You can add up to four elements for 16:9 videos; "Other aspect ratios may have a lower limit."
• Card teasers and watermarks are suppressed while the end screen runs.
• End screens aren't available for videos set as made for kids, and viewers can hide them.
• If an element uses a custom image, YouTube recommends one at least 300 x 300 pixels, and asks you to "consider the video's last 20 seconds when editing it."
• "Your end screen may not always show, or may show differently than designed."

Make the background work with elements where you placed them, moved, or missing entirely (a viewer who hid them, or a device that doesn't show them). A plain branded background with one short line ("Next: the $5 table, part 2") works in all three. YouTube Studio's editor has a grid and snap options for placing elements; place them after the background is final.

Channel intro title cards and other full-screen cards

The intro title card, chapter cards, quote cards and stat cards are all full-frame. Each one stops your footage, so keep it short.

• Intro title card. Put it after a cold open, not before it. The opening seconds belong to the hook; a logo animation in that slot gives a new viewer nothing to stay for. Two to four seconds, one to four words, on a beat of the music or a pause in speech. For music beds, see our guide to AI music for YouTube intros.
• Chapter cards. Short, consistent and fast. Match the wording to your YouTube chapter titles so the timeline and the card agree.
• Quote cards. One quote and a byline. Hold for two read-alouds. If the quote needs three screens, it needs cutting, not more screens.
• Stat cards. One number, one label. The number is the hero, so make it the biggest thing in the frame.

A sample title-card storyboard

Another invented example: an 11-minute Tidewater Workshop episode. The last column shows which elements are overlays (made in the editor) and which are full-frame cards (made in any tool that outputs video).

Tool routes compared

Hand-keyframing means animating position and opacity for every strap, then redoing it whenever a name, a cut or the frame size changes. A template turns that into typing. Descriptions below are per each vendor's site as of October 2026.

Adobe's help page lists "titles, lower thirds, and buttons" among Motion Graphics template uses, and Blackmagic says DaVinci Resolve's "basic title generators let you build simple titles and lower thirds from scratch." Envato Elements lists its lower-third templates for After Effects, Premiere Pro, Apple Motion, Final Cut Pro and DaVinci Resolve.

Our pick for most creators: build straps once in the editor you already use, save them as a template, and spend the effort on cards instead.

Where Oakgen Motion fits, and where it doesn't

Oakgen Motion builds short type-led videos from a description: you say what the piece is for, pick a look, add your brand (read from your website or entered by hand as name, logo and two colours) and optionally screenshots or product shots. Videos made this way are code-built motion graphics rendered from HTML, so the words are real type, not generated pixels. The general strengths and limits are in our overview of what AI motion graphics can and can't do; here's how it maps to a YouTube channel.

What it's good for. The Bold kinetic type look has scene designs that line up with the cards above:

• Slam for an intro title card: a short headline (up to 28 characters) and a small line beneath.
• Statement for chapter cards: a headline and a supporting line.
• Quote for quote cards: up to 110 characters plus a byline.
• Count up for stat cards: a figure and a label.
• Framed screen for a full-frame callout of a screenshot, chart or post you upload.
• Sign-off for an outro sting before your end screen.

For cards you'll cut into an edit, switch the voiceover off and set sound to "None" so you can lay your own audio underneath. Pick 9:16 as well as 16:9 if you make Shorts; each extra size adds to the render.

How you get a card out of it. Lengths are set as a target of 15, 30, 45 or 60 seconds, so you don't make one four-second card on its own. Every Motion video also opens with an Opening scene and closes with an Ending scene (in Bold kinetic, the Slam and Sign-off designs); the storyboard keeps those two pinned first and last. Plan around that: the Slam is your title card, the Sign-off is your outro sting, and the chapter, quote and stat cards go in between. Then slice the MP4 in your editor and drop each card where it belongs. Planning and the storyboard preview use a free daily allowance, and you can edit every word, change a scene's design and reorder scenes there. The exact price shows at the top of the storyboard and on the Make video button, and nothing is charged until you press it. Free accounts have no credits, so rendering needs credits; see pricing. The final download is an MP4 at 1080p or 720p.

Where it doesn't fit:

• Transparent lower thirds, callouts or credits over footage. The output is a full-frame MP4 with no transparent background. Build those in your editor.
• Anything longer than a minute, or 4K.

If you only need a single title card a month, your editor's title tool will do. Use Motion when you need a matching set of cards in both sizes.

Related: AI voice narration for YouTube, the faceless YouTube videos guide, AI videos for YouTube and a product launch video in one evening.

Pre-publish checklist for overlays and cards

• Every overlay is inside a 5% margin from each edge (96 px sides, 54 px top and bottom at 1080p)
• Unlisted test watched paused on a computer, on a phone with controls showing, and with large captions on
• No strap across a face, hands or the product
• Every overlay holds still for twice its read time; first name straps about 4 to 5 seconds
• Exits are faster than entrances
• Text contrast at least 4.5:1 against the brightest shot it appears over
• Two typefaces, fixed colours, one entry direction, fixed positions
• Nothing you want read sits in the last 20 seconds where end screen elements go
• End screen background works with elements moved or hidden
• No more than three flashes in any one second
• 9:16 versions re-laid out, not cropped, and watched on a phone
• Every word proofread in the final export

Frequently asked questions

What is a lower third on YouTube?

It's a short text graphic in the lower part of the frame that names a speaker, labels an object or credits a source. It slides or fades in, holds, then leaves, while the footage keeps playing behind it.

How long should a lower third stay on screen?

Long enough to read it twice once it has stopped moving. For a name and role that usually means four to five seconds the first time and two to three seconds when the person returns.

Where should I put lower thirds so YouTube doesn't cover them?

Inside a 5% margin from every edge, above the area the player controls and captions use, and away from faces. Upload an unlisted test, pause it on a phone and a computer with captions on, and move anything the interface covers.

How long can a YouTube end screen be?

YouTube Help says end screens can be added to the last 5 to 20 seconds of a video, and the video has to be at least 25 seconds long. Plan the last 20 seconds with space left for the elements.

Can I make an animated lower third without After Effects?

Yes. CapCut, Premiere Pro and DaVinci Resolve all ship title or lower-third presets, and browser motion tools such as Jitter can export animation with a transparent background on some plans.

What file format does a transparent overlay need?

One that carries an alpha channel, such as MOV with ProRes 4444 or WebM. A standard MP4 export plays as a full rectangle, so it can only be cut in as a full-screen shot.