Text overlay styles measured

ViewMade ships 7 built-in caption styles for vertical shorts, measured on 22 August 2026 against its own pricing and product documentation. This article describes what those styles are, how they differ from overlay text added in an editor, and where each choice fits a documentary-style short. It is written for people who already publish shorts and are trying to fix format problems, not for people choosing their first tool.

The short answer

A text overlay on a short does two jobs at once: it carries the words when sound is off, and it sets the reading rhythm of the video. ViewMade offers 7 caption styles as of 22 August 2026, all burned into the 1080x1920 frame rather than offered as separate subtitle files. The practical question is not which style looks best in isolation but which one matches your footage type: archive clips, maps, charts, or generated visuals. Word-for-word timing matters more than font choice, because viewers read along with the narration.

Burned-in captions versus separate subtitle files

Captions can live in two places: inside the video frame, or as a sidecar file the platform renders itself. These behave differently, and the difference shows up after upload.

Burned-in captions, which ViewMade produces by default, have three properties worth knowing:

1. They render identically everywhere. YouTube's auto-caption styling, TikTok's caption toggle, and Instagram's native captions cannot move or restyle them.
2. They survive re-encoding. A short downloaded once and re-uploaded to three platforms keeps its exact typography on all three.
3. They cannot be turned off by the viewer, which is a feature for retention-focused publishing and a drawback for accessibility purists who prefer platform-native captions that screen readers can parse.

Sidecar files, such as .srt uploads, let platforms restyle text but give you no control over what the viewer sees. If your channel's look depends on consistent captioning across every clip, burned-in wins. If you rely on accessibility features or want viewers to hide text, sidecar files win. ViewMade embeds subtitles directly into the image, so its output assumes the first position; if your workflow needs editable sidecar files, you would need to add them yourself after download.

What changes between caption styles

The 7 styles ViewMade offers differ along four axes, and knowing the axes helps more than memorizing style names:

Axis	What varies	Why it matters
Position	Where the text block sits in the 1080x1920 frame	Lower-third placement collides with UI overlays on some platforms
Emphasis	Which words get highlighted per line	Highlighted words pace the read against the voiceover
Density	Words per line, lines per screen	Dense blocks slow scanning; sparse ones force faster cuts
Case and weight	Capitalization and boldness	Heavy weights survive compression better

Two of these deserve comment. Position matters because mobile interfaces cover parts of the frame: the right edge hosts engagement icons, and the bottom strip hosts titles and audio metadata on some surfaces. A caption style that hugs the bottom of the frame risks being covered. Density matters because documentary shorts carry factual sentences with numbers, and a viewer pausing to reread a dense line has stopped watching the footage.

We have not measured which axis moves watch time. That claim requires retention data we do not have, so this article treats the axes as design decisions, not performance findings.

Matching overlay style to footage type

The right overlay depends on what is behind the text. This is the section to act on. Take your last published short and classify its footage into one of three buckets, then apply the matching rule:

Archive footage. Historical clips come with their own visual noise: film grain, period graphics, burned-in timestamps from the original broadcast. Text over this needs high contrast and a heavier weight to stay legible. Keep the text block away from the original frame's edges, because archival material is often cropped when reframed to vertical, and edge-hugging text gets cut twice.

Maps and charts. Static graphics leave clean space, usually in the center or on one side. Place text in the empty region instead of over the graphic's labels. A chart with its own axis numbers plus a caption block over it produces two competing text systems, and viewers resolve the conflict by reading neither properly.

Generated or stock visuals. These tend toward smooth gradients and low detail, which tolerate any placement. Here the constraint flips: with nothing competing for attention, the caption becomes the dominant visual element, so density and emphasis choices carry more weight than positioning.

If your shorts mix all three types, as most documentary formats do, pick a single style and accept that it will be suboptimal for one bucket rather than switching styles mid-video. Style switches read as editing errors unless they mark a chapter change.

This is also where ViewMade's approach differs from generation-first tools: it pulls real archive footage, maps, and charts rather than producing frames with a video model, and every render arrives with a media credits file naming each clip's source. The caption system was designed around that footage mix, which is why the style set stays small.

Reading speed and word-for-word timing

Word-for-word timing means each caption line appears synchronized to the spoken words, not in fixed blocks. Two consequences follow for overlay design.

First, timing constrains line length. A line that takes longer to say than to read creates dead air where text sits unchanged; a line that reads faster than speech forces the eye ahead of the audio. Shorter lines timed to phrases keep both channels aligned. When you review a rendered short, watch it muted: if the text still paces sensibly without audio, the timing holds.

Second, timing interacts with emphasis. Highlighting the currently spoken word gives viewers two synchronization cues, audio and visual, and the visual cue works even when sound is off. This is why karaoke-style highlighting appears in most short-form caption systems. The cost is visual busy-ness; on quiet, static shots like maps, constant highlighting can feel restless.

We have not measured whether word-for-word timing improves completion rates compared with block captions. The argument here is mechanical, not statistical: synchronization removes a mismatch between two information channels, and removing mismatches is defensible on principle alone.

What this does not tell you

Three limits apply. First, nothing in this article measures retention: we do not know which of the 7 styles keeps viewers longest, and we will not invent a number. Second, the style count and output specifications reflect ViewMade's documentation as of 22 August 2026; tools change their feature sets, and a style list is a snapshot, not a permanent fact. Third, this article covers caption overlays only. Thumbnail text, title cards, and end-screen text follow different legibility rules because they are viewed at different sizes and durations, and none of the reasoning here transfers automatically to them. If your question is about hook text in the first seconds of a short, this is not that article.

<!-- faq -->

Frequently asked questions

Should captions match my brand font?

Only if the brand font survives vertical rendering and compression. Fonts designed for print often lose weight at small sizes, and thin strokes blur after platform re-encoding. Test one short with the brand font before committing a series to it. Consistency across a channel helps recognition, but an illegible brand font helps nobody. A heavy, simple typeface that approximates the brand's feel usually outperforms the exact font file.

Do I need different caption styles for YouTube and TikTok?

The frame is the same 1080x1920, but the interface furniture differs. TikTok's engagement rail and caption areas occupy predictable zones; YouTube Shorts reserves similar space. One style chosen with conservative margins clears both. Producing two variants per short doubles your render and upload work for a marginal gain, so start with one safe placement and split only if you observe specific coverage problems on one platform.

Can viewers turn off burned-in captions?

No. Once text is part of the image, no player setting removes it. This is deliberate for publishers who want guaranteed caption display, since platform auto-captions are inconsistent in quality and availability. The tradeoff is accessibility: screen readers cannot read pixels. If your audience includes significant assistive-technology use, consider publishing a transcript in the description alongside the burned-in captions.

How many caption styles should a tool offer?

There is no measured answer, and anyone quoting one is guessing. Seven styles, as ViewMade offers, is enough to cover the footage-type buckets described above while keeping the decision fast. A larger library shifts work back onto you: every additional style is another choice to test and another inconsistency risk across a channel. For series publishing, fewer options with clearer guidance beats a wide menu you never finish evaluating.