Captions, Color Basics and Export

The edit is locked. Three jobs remain before anyone sees it: text for people watching with the sound off, a quick pass to make the clips look like they belong together, and a file that plays correctly wherever it lands. None of them is glamorous. All of them show when skipped.

Captions: burned-in or sidecar

A lot of video gets watched on mute, at least for the first seconds. Captions are also how deaf and hard-of-hearing viewers get your story at all. The W3C's accessibility guidance defines captions as a text version of the speech and the non-speech audio needed to understand the content, and WCAG Success Criterion 1.2.2 asks for captions on all prerecorded audio in synchronized media. Caption every video. Not optional.

The choice is how they travel:

My rule: sidecar wherever the platform accepts a file, burned-in where it does not or where people scroll with sound off. YouTube accepts SubRip (.srt) and suggests it, along with SubViewer (.sbv), to people new to caption files, since they hold only timing and text and open in any text editor (as of October 2026). Short vertical feeds are where burned-in captions earn their keep. Some apps also auto-generate captions. Check them: auto captions mangle names and numbers, which are the two things a news story cannot get wrong.

The AV craft chapter Text people can actually read covers caption style in detail: two lines at most, about 40 characters per line, breaks at natural phrases, held long enough to read. It also covers safe areas, which is where burned-in captions go wrong on vertical video: keep them clear of the bottom fifth and right edge, where the app's buttons sit.

1
00:00:00,400 --> 00:00:03,100
My bill went from
48 to 59 dollars.

2
00:00:03,300 --> 00:00:07,800
The Pell Creek board votes
on the rate increase November 12.

AVPrep3 lists transcription, cue editing, SRT import and export and burn-in among its caption features, so you can make one SRT, upload it as a sidecar to YouTube and burn the same cues into the vertical version. Proofread every line against the audio. Spell every name the way its owner spells it.

Color basics: correct, then stop

Color work comes in two stages. Correction makes clips accurate and makes them match. Grading adds a creative look. For journalism, correction is nearly always enough, and a heavy look can make a factual piece feel like an ad.

A correction pass, clip by clip, using your editor's brightness, contrast, saturation and color or white balance controls (VideoStudio has brightness, contrast, saturation, gamma and hue; AVPrep3 has color controls on video clips):

1. Exposure. Is the face bright enough to read, without white blown-out patches? Adjust brightness or exposure until it is.
2. Contrast. Blacks should look black, not gray haze. A little contrast often fixes footage that looks flat.
3. White balance. Find something that should be white or gray: a shirt, a wall, paper. Nudge the color or temperature until it is neutral. Skin should look like skin, not orange and not gray-blue.
4. Match. Put the resident's interview next to her B-roll in the kitchen. Do they look like the same room? Adjust the odd one out.
5. Saturation. Leave it alone, or nudge it a little. Global saturation turns skin orange first.

Judge on more than one screen. Your laptop, your phone, a TV if you have one. The chapter Why your export looks different from your edit covers waveforms, vectorscopes and the levels mismatch behind most washed-out exports, if you want to go further. And remember lesson 3: correction can shift a whole shot, but it cannot fix a face lit by two different colors.

Export settings

The AV craft chapter Getting it out the door has a settings table that works almost everywhere. The short version, cross-checked against YouTube's own recommended upload settings (as of October 2026):

Why upload above the platform's own number? Every platform re-encodes your upload. Giving its encoder a clean, generous source means less damage. A file that is already squeezed hard looks worse after the platform squeezes it again.

Frame rate matching deserves the emphasis. YouTube's guidance says to encode at the frame rate you recorded. Exporting 30 fps footage at 60 does not make it smoother; it just repeats frames and doubles the data. Exporting it at 24 drops frames and makes motion judder.

When there is a size limit

Email, messaging apps and some upload forms have hard caps. Personal Gmail attachments, for instance, top out at 25 MB (as of October 2026). File size is bitrate times duration, so you can work the number out:

Rather than doing that math by hand, use StudioConverter. It compresses to a target file size in MB, with optional two-pass encoding for closer accuracy, and can cap resolution and frame rate. Its loudness tool sets a LUFS target with a peak limit if the audio needs it. AVPrep3's export dialog also offers a target file size and platform presets.

If the result looks bad, the fixes in order of least damage: shorten the video, drop to 720p, use two-pass. The AV chapter notes that 720p at a low bitrate usually looks better than 1080p at the same bitrate, because the encoder spreads the same data over fewer pixels.

Before you publish

Watch the exported file, start to finish, not the timeline. Check the first five seconds and the last two (exporters love to clip the ending). Read the captions on a phone. Listen on the phone speaker at half volume. Then save three things in the project folder: the project file, the source media, and one high-quality export without platform crops or burned-in captions. That last one is your archive master; you will want it the day someone asks for a different version.