Captions for Family Videos That Don't Cover Faces

You found the photo: Grandpa at the 1987 family reunion, laughing, holding the baby who is now your cousin with three kids of her own. You type “Reunión familiar en San Antonio” across the bottom, export, and send it. On the small screen, the caption sits right across the baby’s face. On a second photo, it lands on Grandma’s mouth mid-word.

Captions for a family video do two jobs: they tell the viewer who, where and when, and they stay out of the way of the faces that are the whole point. The good news is that the rules for doing both are well worked out. Broadcast caption standards, accessibility guidelines and eye-tracking research all point to the same short list: keep text short, high-contrast, away from eyes and mouths, on screen long enough to read, and checked on the real screen.

Key Takeaways

• In free viewing, people looked at faces about 16.6 times more and at text about 11.1 times more than at comparable regions of an image (Cerf, Frady & Koch, 2009). Put a caption on a face and you make two magnets fight.
• US broadcast caption rules say captions must not block “character faces” or featured text, and must be sized for legibility (47 CFR 79.1).
• Accessibility guidelines ask for at least 4.5:1 contrast between text and background, partly because contrast sensitivity drops with age (W3C WCAG 2.2).
• Viewers in an eye-tracking study kept up with both the text and the images even at fast subtitle speeds, but they spent half or more of their time reading (Szarkowska & Gerber-Morón, 2018). Every word you add takes time away from the faces.
• Practical floor: one or two short lines, moved to the top when faces sit low, held long enough to read at about 130 to 160 words per minute.

Why captions keep landing on the wrong part of the photo

People who make tribute videos describe the same problems: “the text covers their faces,” “the caption was unreadable on the TV,” “it went by too fast for Grandma to read.” Most editors drop titles in the same default spot, centered near the bottom, and most family photos put faces right where that default lands. Group photos are worse. Ten people standing in a row fill the lower half of the frame, and a lower-third caption sits across their chins.

Then there is the screen. You add captions on a laptop, where a 24-point title looks big. The finished video plays on a phone, a tablet or a small display, and the same title shrinks into a line of gray crumbs. The opposite also happens: text sized for a phone fills a third of a TV.

And there is the reading. A caption you wrote is a caption you can read in a glance, because you already know what it says. A first-time viewer has to read it while also looking at the photo, and if it disappears after two seconds they pick one. They usually pick the face.

What the research says about captions and faces

Faces and text both pull the eye

Moran Cerf, Paxon Frady and Christof Koch tracked where people looked at images in the Journal of Vision (2009). In free viewing, participants looked at faces about 16.6 times more, and at text about 11.1 times more, than at comparable regions of the image matched for size and position. Text was nearly as strong a magnet as a face. The title of the paper states the broader claim: the pull held “independent of the task.”

For a caption maker, the implication is simple. A caption is not a neutral label. It competes for the same glances as the faces. Put it on top of a face and the viewer either reads the caption and misses the expression, or looks at the expression and misses the caption.

People can read and watch, but reading takes the time

Agnieszka Szarkowska and Olivia Gerber-Morón tested subtitle speeds of 12, 16 and 20 characters per second in PLoS ONE (2018). Most viewers read the subtitles and followed the images, even at the fastest speed. The cost showed up in where their eyes went: in one experiment viewers spent about half their time on subtitles at the slowest speed and about two-thirds at the fastest. Slow subtitles also caused more re-reading, which frustrated people and lowered enjoyment.

Their participants were adults watching film clips, not family photos, so treat the numbers as direction rather than rule. The direction is useful: more text means more time on text, and text that sits too long gets re-read. For a memory video, the goal is the shortest caption that still says who and where, timed to match.

How long a caption should stay up

The Described and Captioned Media Program (DCMP) publishes a Captioning Key used for educational media. For younger audiences it recommends a presentation rate not exceeding 130 words per minute, with 140 for middle levels and 160 for upper levels. At 130 words per minute, a ten-word caption needs about 4.6 seconds; at 160, about 3.8 seconds. If grandparents or kids will watch, use the slower number.

What broadcast rules say about placement and contrast

US television caption rules (47 CFR 79.1(j)(2)) require captions to be viewable without blocking “other important visual content on the screen,” naming “character faces” and featured text such as graphics and credits. They also require caption font sized for legibility and captions that stay on the screen rather than running off the edge.

For contrast, the W3C’s Web Content Accessibility Guidelines (WCAG 2.2) ask for at least 4.5:1 between normal text and its background, and 3:1 for large text. The guideline explains that the 4.5:1 ratio compensates for the loss of contrast sensitivity typical of about 20/40 vision, which it notes is common around age 80. White text on a sky, a white wall or a wedding dress fails that test even when it looks fine to you.

Caption placement decision table

Use the table when you place each caption.

Where the faces are	Where the caption goes	Watch out for	Basis
Faces in the upper half (portrait, close-up)	Bottom of the frame	Hands, a cake, a baby on a lap	47 CFR 79.1 placement rule
Faces in the lower half (group photo, kids on the floor)	Top of the frame	Heads near the top edge	47 CFR 79.1; broadcast practice
Faces fill the frame (tight close-up)	Before or after the shot, or on a plain frame between photos	Cutting the caption into chins	Cerf et al. 2009 (faces and text compete)
Busy, bright background (sky, snow, white dress)	Text with a soft shadow or a semi-transparent band	Thin white letters with no backing	WCAG 2.2 contrast 4.5:1
Long caption (more than about 10 words)	Split into two captions or shorten	Text still up when the photo changes	DCMP presentation rate; Szarkowska & Gerber-Morón 2018
Video will play on a small screen	Larger text, fewer words	Detail that disappears below a certain size	47 CFR 79.1 legibility; test on device

The pattern is the same in every row: the faces decide where the words go, not the editor’s default.

How to add captions without covering faces

Step 1: Decide which photos need a caption at all

Not every photo needs words. A caption earns its place when the viewer would otherwise ask “who is that?” or “where was this?” A photo of Mom laughing in the kitchen does not need “Mom laughing in the kitchen.” Pick the photos where a name, a place or a year adds something, and leave the rest clean. Common mistake: captioning every photo, which trains viewers to read instead of look.

Step 2: Write the shortest caption that answers who, where or when

Aim for four to eight words. “Abuela Rosa, Monterrey, 1974” beats “This is a photo of Abuela Rosa when she lived in Monterrey back in 1974.” Names and places carry the meaning; filler words cost reading time. If a story needs more, split it across two photos or put a short title card between them. Common mistake: writing the caption like a sentence in a letter.

Step 3: Look at where the faces are before choosing a position

For each photo, find the eyes and mouths. If they are in the top half, the caption goes at the bottom. If they are in the bottom half, as in most group shots, move it to the top. If faces fill the frame, put the caption on the frame before or after, or on a plain title card. Most editors let you drag a title anywhere or pick a top or bottom style. Common mistake: accepting the editor’s default lower-third position for every photo.

Step 4: Make the text readable against the background

Check the contrast. On bright or busy areas, add a soft drop shadow, an outline or a semi-transparent dark band behind the text. A simple, solid sans-serif font at a regular or bold weight reads better on small screens than a thin script. The DCMP Captioning Key also prefers mixed case to all caps for readability. Common mistake: choosing a handwritten font that looks lovely on a laptop and turns to mush on a phone.

Step 5: Hold each caption long enough to read

Use the DCMP rates as a guide: roughly 130 words per minute for slower readers, 160 for practiced ones. A six-word caption needs about two and a half to three seconds; a ten-word caption needs about four to five. If the photo is on screen for less time than that, shorten the caption or lengthen the shot. Common mistake: letting the caption fade in late and leave early, which halves its reading time.

Step 6: Test on the screen the video will play on

Export and play the video on the actual device: the phone, the tablet, the small TV, the screen in the living room. Hold it at the distance the recipient will. Ask someone who has not seen the video to read each caption aloud. If they squint, lean in or miss one, make the text bigger, shorter or longer on screen. Common mistake: approving the captions on the editing screen only.

What not to do

Don’t put a caption across eyes or mouths, even briefly. Expressions are why the photo is there. And don’t fill the frame with dates on every photo; a single title card that says “1970s” before a run of photos does the job of twenty captions.

How to make it personal

If you have 10 minutes

Open your three favorite photos and, for each one, write the caption only you could write: not “Grandpa and me,” but “Grandpa teaching me to fish, Lake Texoma.” Keep each under eight words. That list is half your captions done.

If you have a weekend

Call or text the oldest relative you can reach and ask them to name the people and places in the photos you are unsure about. Write down exactly how they say the names, including nicknames and accents. A caption with “Tío Chuy” spelled the way the family says it is worth more than a perfect font. Then check the spelling with the same care you would give a printed gift; here is a checklist for names and dates.

If you want it to last years

Keep a plain text file next to the video that lists every photo, who is in it, where and when. Captions burned into a video can’t be searched or corrected; a list can. Future you, and future cousins, will use that list long after the video is old, which is the same logic behind turning the camera roll’s best moments into something real.

The question to ask yourself: “If someone could read only the caption or see only the face, which one would I want them to keep?”

Frequently asked questions

Where should captions go on a family video?

Usually at the bottom, unless faces are there. In group photos and shots where people sit low in the frame, move captions to the top. US broadcast caption rules say captions must not block faces or featured text, and the same logic applies at home: the faces decide where the words go.

How long should a caption stay on screen?

Long enough to read twice comfortably. The DCMP Captioning Key recommends rates of about 130 to 160 words per minute depending on the audience, so a six-word caption needs roughly two and a half to three seconds and a ten-word caption about four to five. Use the slower rate if older relatives or kids will watch.

What font and color are easiest to read in a video?

A simple sans-serif font at regular or bold weight, in mixed case, with strong contrast against the background. Accessibility guidelines ask for at least 4.5:1 contrast. White text with a soft shadow or a semi-transparent dark band behind it works on most photos.

Should every photo in a memory video have a caption?

No. Caption the photos where a name, place or year adds something the viewer can’t see. Eye-tracking research shows text pulls attention almost as strongly as faces, so captions on every photo turn watching into reading. A single title card before a group of photos often does more.

Why are my captions unreadable on a phone or small screen?

Text sized on a laptop shrinks on a phone, thin fonts lose strokes, and light text disappears on bright backgrounds. Make the text larger, use fewer words, add a shadow or band, and always test on the device the video will play on.

---

About the author

Ricky Flores is the founder of Watch N Wow and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. In Dallas, Texas, he builds the Mini Memory TV by hand: a small retro TV that plays a video, a message and an original song made from your photos and your story. Read more at watchnwow.com.

---

Sources

1. Cerf, M., Frady, E. P., & Koch, C. (2009). “Faces and text attract gaze independent of the task: Experimental data and computer model.” Journal of Vision, 9(12), 10. https://doi.org/10.1167/9.12.10
2. Szarkowska, A., & Gerber-Morón, O. (2018). “Viewers can keep up with fast subtitles: Evidence from eye movements.” PLoS ONE, 13(6), e0199331. https://doi.org/10.1371/journal.pone.0199331
3. Described and Captioned Media Program. “Captioning Key: Presentation Rate.” https://dcmp.org/learn/601
4. US Code of Federal Regulations, 47 CFR § 79.1 (j)(2), caption quality standards. Legal Information Institute, Cornell Law School. https://www.law.cornell.edu/cfr/text/47/79.1
5. W3C. “Understanding Success Criterion 1.4.3: Contrast (Minimum).” Web Content Accessibility Guidelines 2.2. https://www.w3.org/WAI/WCAG22/Understanding/contrast-minimum.html
6. Described and Captioned Media Program. “Captioning Key.” https://dcmp.org/learn/captioningkey