Wedding Video Subtitles: Vows, Toasts and Bilingual Ceremonies, by the Broadcasters' Rules

Subtitle a wedding film the way broadcasters subtitle theirs: at most two lines of about 42 characters, each on screen for between five-sixths of a second and seven seconds, at no more than 20 characters a second (Netflix) or 160 to 180 words a minute (the BBC). Keep vows and toasts word for word. For a ceremony held partly in another language, burn the translation into the picture, so it shows on every player. Then deliver the file each player can read: SRT or WebVTT in UTF-8 for Vimeo and YouTube, and burned-in text for a file played from a USB stick.

The rules below are Netflix's and the BBC's, from their published style guides; what they mean for vows, toasts and two-language ceremonies is our reading, labelled as such. The formats each platform accepts are from Vimeo's, YouTube's and one TV maker's own help pages, and what each editor can do from Adobe's, Blackmagic Design's and Apple's documentation.

The broadcasters' rules, applied to vows and toasts

• Lines. Netflix allows 42 characters a line and two lines at most, kept to one line when it fits. Break after punctuation or before a conjunction or preposition, never between a first name and a last name, and prefer the longer line at the bottom.
• Duration. Each subtitle stays up for at least five-sixths of a second, 20 frames at 24 fps, and at most 7 seconds.
• Reading speed. Netflix caps adult programs at 20 characters a second. The BBC recommends 160 to 180 words a minute, 0.33 to 0.375 seconds a word, and as a floor about 0.3 seconds a word: 1.2 seconds for a four-word subtitle.
• Verbatim. Netflix's subtitles for deaf and hard-of-hearing viewers keep transcriptions as close to the audio as possible and do not paraphrase. The BBC says viewers tend to prefer verbatim subtitles, and that words a viewer can lip-read in a close-up should not be edited out.
• Lips moving, words up. The BBC wants a subtitle on screen whenever someone's lips are moving, and none left up after the speaker has gone.

Our arithmetic at Netflix's 20 characters a second: a full two-line subtitle of 84 characters needs at least 4.2 seconds, a full single line 2.1. The BBC subtitles most programs word for word up to about 160 to 180 words a minute and expects some editing above that, so a speaker faster than that is where the choice comes up.

On our reading, vows are never condensed: they are the words the couple will most want exactly, and they are said in close-up, which is where the BBC warns against cutting lip-readable words. For a fast toast, use the BBC's own allowance: one speaker's subtitles may slip slightly out of sync during a run of subtitles as long as they are back in sync by the end of it. Keep every word, let the run lag a little, and bring it back at the pause.

Sound labels are worth borrowing from Netflix's deaf and hard-of-hearing rules too: [laughter], [applause] or [music playing] in square brackets, lowercase, only where the picture does not already show it, and a name label such as [Maria] only when the speaker is off screen. A toast heard over a wide shot of the room is the case at a wedding.

A ceremony in two languages

Netflix's guides deal in three kinds of timed text, and the third is the useful one here. Subtitles for deaf and hard-of-hearing viewers transcribe the same language, with sound labels. Subtitles translate a language the viewer does not speak. A forced narrative is a text overlay for "alternate languages meant to be understood by the viewer" (Netflix): Netflix's player shows it while subtitles are switched off, and because it then hides it when a full subtitle file is switched on, Netflix requires every forced line to appear in every full file as well.

At a wedding everything said in either language was meant to be understood by someone in the family, so on our reading the translation is a forced narrative: it belongs in the film, not behind a menu. Netflix can deliver it as a separate file because its own player knows to show it. Vimeo's upload offers two types, subtitle and caption, and a TV playing a downloaded file offers whatever its maker built, so for a wedding film the way to get the same result on every screen is to burn the translation into the picture. Keep a same-language file optional for anyone who wants every word, and leave the burned-in lines out of it so the two never stack. If both families need translating in both directions, deliver two versions rather than four lines of text at once.

• Ask the couple, before the wedding, which parts will be in which language and which language each version of the film should be subtitled in.
• Spell every name the couple's way. Netflix keeps accents and diacritics in names, Mónica not Monica, and so should a wedding film.
• Have the translation checked by someone who speaks both languages, ideally from the family. Readings, blessings and religious rites are the parts to have checked first.
• Keep subtitles off your on-screen titles. Netflix places subtitles at the top or the bottom of the screen to avoid covering on-screen text, and the same goes for a chapter title in a doc edit.

Which file plays where

• Vimeo: upload an SRT or WebVTT file in the video's Languages panel and choose whether it is a subtitle or a caption. Vimeo recommends WebVTT and requires UTF-8 encoding so special characters display correctly. Automatic subtitles and AI translation of subtitles are on Free and paid plans (Vimeo Help).
• YouTube: plain SRT or SubViewer files in UTF-8, with no styling recognized; WebVTT or TTML when you want styling and position (YouTube Help).
• A file played from a USB stick on a TV: Philips says its sets read external SRT, SUB and SAMI subtitles only from a USB drive, and only in the Latin-1 Western European character set (Philips support). On those sets, Hindi, Punjabi, Arabic, Hebrew, Chinese or Korean subtitles will not show, and nor will a Polish "ł". Other makers differ, which is the point: you cannot know the couple's TV, so burn in anything that must be read.
• A vertical teaser: the BBC's guide puts the equivalent of its 37 characters at about 25 characters across 90% of a 9:16 frame, so break social cut-downs into shorter lines, and burn them in.

Where the film is delivered, and how, is in the delivery guide. On our reading the safe default for a wedding film is burned-in text for anything that must be read, such as a translation, and a separate file for anything optional.

Making the file in your editor

All three editors can turn speech into subtitles, with differences that matter at a wedding. From each maker's own documentation, read on November 3, 2026:

• Premiere Pro: Speech to Text is included in the subscription and works offline once a language pack is installed. It covers 18 languages, Spanish, Hindi, Japanese, Korean, Mandarin and Cantonese among them, and not, for example, Punjabi, Urdu, Arabic or Hebrew. Captions export as SRT, as broadcast formats (SCC, MCC, STL and TTML), embedded in the file or burned in; WebVTT is not on Adobe's list of caption formats. Translating captions goes through a cloud service using Google Translate and Microsoft Translator. Adobe also says to make captions from the finished sequence, not while editing by text.
• DaVinci Resolve 21: transcription and Create Subtitles from Audio are Studio Version Only. Fourteen languages are built in; an extra download adds 43 more, Arabic, Hebrew, Hindi, Urdu, Polish, Tagalog and Vietnamese among them, which the project settings call beta. Subtitles export as SRT or WebVTT, render as IMSC1, DFXP, SRT or WebVTT files, embed as CEA-608, or burn in. Blackmagic warns that automatic subtitles still commonly fail where dialogue clips overlap on several tracks, which describes a synced wedding timeline: mute every track but the one you want transcribed.
• Final Cut Pro 12.4: Generate Captions (Transcribe to Captions when it arrived in 11.0) needs a Mac with Apple silicon and is "available in U.S. English only". It makes iTT captions, which convert to SRT or CEA-608; only CEA-608 embeds in the exported file, SRT and iTT go out as separate files, and several share destinations can burn captions in.

For a ceremony in two languages that list is the first thing to check. A Spanish and English wedding transcribes in full in Premiere Pro or Resolve Studio but not in Final Cut Pro; Hebrew or Urdu only in Resolve Studio's beta download; Punjabi in none of the three, by their own lists. Whatever transcribes it, a person who speaks both languages checks the result.

Sync first (how to sync multicam wedding footage), then transcribe the clean recording made at the source, the lav or the board feed, rather than a camera's own microphone across the room: the clearer the words, the fewer the corrections. The transcript has a second use before it becomes subtitles. It is the fastest way to find soundbites, because you can read a toast in a minute and listen only once you have found the line.

Check before you deliver

• Names, places and family terms spelled the couple's way, with their accents.
• No line over 42 characters, no subtitle over two lines.
• Nothing on screen for less than five-sixths of a second or more than 7.
• No run faster than 20 characters a second, apart from a fast speaker you have chosen to let lag and catch up.
• Vows complete, word for word.
• The file saved as UTF-8, and a test play on the platform, on a phone and on a TV.

Where Pixally Prep fits

Pixally Prep transcribes the speech in your footage and writes a subtitle file for each timeline it builds, so the transcript and the timings exist before you open your editor. It also pulls soundbites aside for the highlight, each one a whole sentence with every angle under it, and builds the first cut of each doc edit you choose. The first wedding is free, one per person and one per computer, and each one after that is $8.99. Check its subtitles against the list above like any other.