Forced narrative vs SDH vs CC
Forced narrative, SDH, and CC are three caption jobs. Mixing them is how you ship the wrong file to a platform.
Name the job before you name the file. Forced narrative, SDH, and CC are three different caption jobs. Teams mix them, export one sidecar, and then watch a platform show foreign signs with no dialogue, or dump every sound effect onto a TikTok burn that only needed a hook line. The fix is vocabulary, then pipeline.
Forced narrative is selective, not shy
Forced narrative (often called forced narrative subs) covers the bits a viewer must understand even if they skip full subtitles: a shop sign in another language, a critical phone text, a plot-essential line spoken off-screen in a language the master track does not carry for that audience. The file is sparse on purpose.
If you ship a full dialogue subtitle file labeled as forced narrative, players that honor forced tracks will spam the screen. If you ship nothing, international viewers miss the sign that explains the joke. The middle path is a dedicated forced list timed only to those essential events.
Social clips rarely need classic theatrical forced tracks, but the mindset still helps. Ask: what must be readable even if the viewer never opens a full CC track? For many UGC ads, that answer is the hook line itself, which is why people burn it. For narrative shorts with in-world text, you may still want a forced pass separate from marketing captions.
For a deeper split between sparse and full files, see forced narrative subs versus full subtitle files. Keep the filenames honest so the next editor does not "help" by merging everything into one SRT.
Forced narrative is also where translators and editors should talk early. A sign that is plot-critical in Japanese may need a short English forced line even when the rest of the cut stays in English audio. If that line is buried in a full CC file that users disable, the story breaks. Sparse forced tracks exist to survive "subtitles off."
SDH carries sound for people who need it
SDH means subtitles for the deaf and hard of hearing. These tracks include dialogue plus meaningful non-speech audio: [music], [door slams], speaker IDs when the face is off camera. SDH is an access product. It is denser than a marketing karaoke burn and denser than forced narrative.
Dumping SDH into a TikTok-safe karaoke style usually fails. The screen fills with brackets and cues the mute-feed viewer did not ask for, while the viewer who needed SDH still lacks proper speaker labeling and consistent cues. Different jobs.
If your brand ships long-form YouTube with accessibility commitments, budget an SDH pass. If you only ship mute-first ads, you may still burn dialogue hooks without claiming SDH completeness. Do not label a two-word marketing burn as SDH. That mislabels the file and misleads your compliance checklist.
Quality SDH has house rules: how you label music, how you handle overlapping speech, when you identify speakers. Without rules, two freelancers produce two dialects of brackets. Viewers feel the inconsistency even if they cannot name it. Write a one-page SDH guide before you scale volume.
CC is the broad label people misuse
CC (closed captions) often means user-togglable captions authored as a separate track. In casual speech, people say CC for any on-screen words. In production, closed versus burned-in is a delivery distinction. Closed captions can be turned off. Burned-in captions are in the pixels.
Platforms blur the language further. A "CC" button may point to auto-captions, uploaded SRT, or creator-edited captions. Your job is to know which deliverable you produced. For mute-first social, burned dialogue lines are usually the product. For accessibility on long-form, a proper closed track (often SDH-quality) is the product.
Mixing names in a spreadsheet causes wrong uploads: an SDH file attached where a forced track was expected, or a marketing burn uploaded as if it satisfied an accessibility contract. Write the job on the ticket: forced, SDH, marketing burn, or full dialogue CC.
Auto-captions are a draft, not a deliverable category. They can seed timing for a marketing burn or a CC edit. They do not automatically become SDH. Cleaning auto output into SDH is real work with real minutes on the clock. Budget those minutes or you will ship brackets that lie about sounds that never happened.
Pick the pipeline that matches the platform
For TikTok and Reels ads, start with a marketing caption burn sized for safe area, then decide if you also need a downloadable track. Use an AI caption generator to draft dialogue timing, then edit down for the job. Karaoke hooks stay short. SDH stays complete. Forced stays sparse.
For films and long-form, keep separate files with clear names: film_forced.srt, film_sdh.srt, film_en.cc.srt. Never overwrite one with another. Players and platforms map tracks by metadata and filename conventions your distributor documents.
Emotionally, this work is kindness disguised as pedantry. Viewers who need SDH feel respected when cues are real. Viewers on mute feel respected when the hook is readable. Viewers watching foreign signage feel respected when forced lines appear and then get out of the way.
A practical weekly habit: in standup, say the caption job out loud for each deliverable. "This one is marketing burn only." "This one needs SDH for YouTube." "This one needs forced for the French sign." Thirty seconds of naming prevents Friday file swaps.
Ship the right file. If you need burns for social from the same studio that generates the clip, sign up at Versely and label the caption job on the brief before anyone hits export.
Distributors and platform uploaders often have their own droplists for caption roles. Match your internal job names to their droplist labels before the night-of-delivery scramble. If their UI only offers "subtitles" and "closed captions," put the mapping in the delivery memo so nobody improvises. Improvisation is how forced narrative gets uploaded as full CC and then looks "broken" in QA.
Educate stakeholders with one sentence each. Forced: only what you must read with subs off. SDH: dialogue plus meaningful sounds for access. CC: toggleable captions, not automatically burned. Marketing burn: pixels for mute feeds. Four sentences prevent weeks of Slack archaeology.
FAQ
Can one SRT cover forced narrative and SDH?
It should not. Forced tracks are sparse by design. SDH tracks are dense and include non-speech cues. Merging them produces the wrong experience in players that expect a true forced track.
Are burned TikTok captions the same as CC?
No. Burned captions are in the video frames. Closed captions are a separable track the viewer can toggle. Social burns solve mute viewing. CC/SDH solve toggleable access, especially on long-form.
Which caption job should UGC ads prioritize?
Usually a marketing burn of the spoken hook, timed short and placed in safe area. Add SDH or forced deliverables when the channel or story truly requires them, and name those files separately.