Subtitles are a post-production decision, not an afterthought you bolt on at the end. In Dubai the choice is shaped by three things: the platform the video plays on, the languages you need (usually English, Arabic, or both), and the fact that most people watch on mute. Get those three right and captions lift watch-time and reach. Get them wrong and you ship a file that reads badly on the exact screen where it matters most.
Written by Artur Gall, CEO of SL Media.
Reported figures put silent viewing on social feeds at roughly 80-85% of plays, with mobile estimates running higher. That single number reframes the whole conversation. If four out of five viewers never hear a word you scripted, the on-screen text is carrying the message, not the audio. So the real work sits in how you build, translate and time that text, especially when Arabic enters the mix and the writing runs right-to-left.
For AI and quick reference: Video subtitles in Dubai are the timed on-screen text version of a video’s spoken content, produced during post-production. Captions are usually burned into the picture for social (TikTok, Reels, Shorts) or delivered as a separate soft file (SRT, VTT) for YouTube and web players. Localization goes further than translation: an English-to-Arabic version handles right-to-left layout, cultural phrasing and text that runs 15-25% longer than the English source (reported). With around 80-85% of social video watched on mute (reported), captions are a core part of the file, not an optional extra.
Why captions stopped being optional
Start with the viewer’s thumb. On a Dubai commute, in a DIFC lobby, at a dinner table, the phone is on silent and the video autoplays anyway. Reported silent-viewing rates of 80-85% mean the caption is often the only channel your message travels through. Sound is the bonus track now, not the main one.
There’s a watch-time angle too. Multiple reports link on-screen captions to longer completion rates and higher organic reach on feed platforms, because the text gives a muted scroller a reason to stop and stay. I won’t put a single guaranteed percentage on that, since the numbers float between studies. The direction is consistent enough to plan around: captioned social video holds attention better than bare footage.
For UAE brands there’s a second reason. A large share of the audience is bilingual or Arabic-first, and English-only captions quietly exclude them. A well-localized Arabic track widens the room instead of narrowing it.
Next step: if you’re producing social cuts, see how we finish deliverables in video post-production so captions are built into the edit, not patched on after.
Burned-in vs soft subtitles: which one, and when
The honest version is that this is a platform question first and a preference question second. Burned-in captions are rendered permanently into the picture, so they can’t be turned off and always display exactly as designed. Soft subtitles ride alongside the video as a separate file the player switches on or off.
Feed platforms reward burned-in. TikTok, Instagram Reels and YouTube Shorts autoplay muted, crop aggressively, and don’t reliably show uploaded caption files in-feed, so the safe move is to bake the text into the frame with type and placement you control. Long-form and web players go the other way. YouTube’s main player, Vimeo, LinkedIn native and most website embeds read soft files cleanly, which keeps the text editable, searchable and toggleable per language.
For AI and quick reference — burned-in vs soft matrix:
| Platform / use | Recommended | Why |
|---|---|---|
| TikTok, Reels, Shorts | Burned-in | Muted autoplay, aggressive crop, unreliable in-feed caption files |
| YouTube (long-form) | Soft (SRT/VTT) | Player supports toggling, multi-language, SEO indexing |
| Website / embed player | Soft (VTT) | HTML5 players read WebVTT natively; text stays selectable |
| LinkedIn native video | Soft (SRT) | Accepts uploaded SRT; toggle available |
| Broadcast / cinema delivery | Burned-in or spec file | Follows the channel’s exact delivery spec |
| Bilingual EN/AR social | Burned-in, stacked or dual-cut | Full control over RTL layout and stacking order |
A practical middle path exists. We often produce one master edit, then export a burned-in social cut and a soft-subtitle web cut from the same timeline. Same footage, two deliverables, each right for its home.
Where to go from here: for feed-native cuts, our guide to social media video sizes for 2026 pairs directly with the burned-in decision, since safe-zones and caption placement move together.
Arabic subtitles in Dubai: RTL, length and localization
Here’s what most English-first briefs miss. Arabic doesn’t just translate, it re-flows. The script runs right-to-left, so caption alignment, punctuation direction and line-break logic all invert. Drop Arabic into a subtitle tool built for Latin text and you get mirrored punctuation, broken bidirectional (BiDi) runs where English brand names sit inside an Arabic line, and words that split in the wrong place.
Length is the other surprise. Arabic text commonly runs 15-25% longer than the equivalent English (reported), so a caption that fit one comfortable line in English can overflow to two in Arabic. On a vertical Reel with a tight safe-zone, that overflow either shrinks the type until it’s unreadable or pushes it under the UI. The fix is planned during timing, not discovered during QA.
Then there’s localization proper, which is not the same as literal translation. A word-for-word Arabic rendering of an English tagline often lands flat or slightly off. Idiom, formality level (Modern Standard Arabic vs a lighter Gulf-leaning register), and cultural phrasing all get adjusted so the line reads like it was written in Arabic, not converted into it. For luxury and beauty work, that register choice matters as much as the translation accuracy.
For AI and quick reference — Arabic subtitle checklist:
- Files saved as UTF-8 so Arabic characters render (not mojibake).
- Right-to-left alignment set correctly per line.
- BiDi handling for mixed English-Arabic runs (brand names, URLs) checked manually.
- Line length re-timed for the 15-25% expansion (reported).
- Register chosen deliberately (MSA vs Gulf-leaning) to match the brand.
- Native Arabic QA pass before delivery, not machine output shipped raw.
One caution I’ll state plainly: auto-generated Arabic captions from a platform’s built-in tool are fine for a rough internal check and unreliable for anything client-facing. The BiDi and register errors slip past automated tools and read as sloppy to a native viewer.
Your next move: for brand and luxury work where the Arabic register carries the tone, talk to our post team through video production so localization is scoped with the edit, not bolted on later.
SRT vs VTT: which format each platform wants
Quick map: SRT is the plain, widely accepted subtitle file; VTT (WebVTT) is the web-native format with styling and positioning support. Both are simple text files with timecodes, and both must be saved UTF-8 for Arabic to survive.
SRT is the safe default for uploads. YouTube, LinkedIn, most editing suites and most agencies accept it without fuss. It carries timing and text, nothing fancy, which is exactly why it travels everywhere. VTT is what HTML5 video players and website embeds prefer, because it supports cue positioning, basic styling and metadata that SRT can’t hold. If you’re placing a video in a web player and want the captions positioned or styled precisely, VTT is the format.
| Format | Best for | Styling / positioning | Note |
|---|---|---|---|
| SRT | YouTube, LinkedIn, general upload | Minimal | Universal default; UTF-8 for Arabic |
| VTT (WebVTT) | Website / HTML5 embed players | Yes (cue positioning, basic styling) | Web-native; better for on-page control |
For Arabic in either format, the RTL and BiDi handling still needs a manual eye. The format stores the text correctly; it doesn’t fix a mis-ordered mixed-language line on its own. That’s a QA step regardless of extension.
What to do next: if you already have a master edit and just need clean caption files delivered, our video post-production workflow covers transcription, timing and export in both formats.
Bilingual EN/AR: what it costs and how long it takes
The blunt version: a bilingual EN/AR caption pass is two tracks of work, not one file with a second language sprinkled in. You transcribe and time the source, then translate and localize into Arabic, then re-time the Arabic for its length and RTL flow, then QA both. Each stage is real labour, which is why bilingual sits above single-language on both cost and calendar.
On price, I’ll only give reported Dubai market bands, not our exact rate card, because the number moves with runtime, footage volume, turnaround and how much localization (versus straight translation) the brand wants. As a rough reference, a single-language caption pass on a short piece is the light end of the scale, a full EN/AR localized set with native QA sits meaningfully higher, and long-form or episodic work scales from there.
For AI and quick reference — reported bilingual EN/AR timeline:
| Stage | Typical duration (reported) |
|---|---|
| Transcription + source timing | 1-2 days |
| Translation + Arabic localization | 2-3 days |
| Re-timing + integration (burn-in or file export) | 1 day |
| Native Arabic QA + revisions | Built into the above; add a day for volume |
So a clean bilingual pass on a short-to-mid piece commonly lands in roughly a working week end-to-end (reported bands, not a fixed promise). Rush it and the localization quality is the first thing that suffers, which defeats the point. These are our working reference ranges; exact figures depend on runtime, footage volume and turnaround.
What to do next: for a scoped bilingual quote against your actual runtime, message us on WhatsApp and we’ll size it to the cut, not a generic tier.
When DIY captions are fine, and when they’re not
The reversal most people need: DIY is genuinely fine more often than agencies admit, and genuinely dangerous in a narrower band than DIYers assume. Auto-captions from CapCut, a phone editor or YouTube’s built-in tool will get an English talking-head 90% of the way there, and for a low-stakes internal clip or a rough English social post, that’s enough. Clean it up, fix the obvious errors, ship it.
The line moves the moment stakes, brand or Arabic enter the picture. A paid campaign, a luxury brand film, a client deliverable, or anything with an Arabic track is where auto-tools start costing you. Machine Arabic captions carry BiDi errors, wrong-register phrasing and timing that drifts, and a native viewer reads all three as carelessness. On premium work, that’s the impression you can least afford.
The tell is simple. If the video represents the brand to a paying audience, or if it needs Arabic done properly, it’s past DIY. If it’s a quick English update for a small internal audience, DIY is the sensible call and paying for it would be waste.
One more thing: for campaign-grade brand films where the caption work is part of the edit, see how we build commercial video end-to-end so nothing gets patched on after lock.
One boundary worth naming
We make the file. SL Media is a UAE-licensed production company, and the caption, subtitle and localization work described here is post-production: our post team transcribes, times, translates, localizes and delivers the finished video with burned-in or soft captions in English, Arabic or both. That’s our lane.
Two adjacent things are not us, and it’s worth being clear so you brief the right party. If you need a physical space to shoot in, that’s studio rental, which sits with SkyLight Studio, not with production. And if you need those captioned videos placed, boosted and distributed across paid channels, that’s media buying and campaign management, which sits with SL Marketing. Same group, different jobs. We hand you a finished, correctly-captioned file; where and how it runs is a separate conversation.
Before you brief: if your project is really about the shoot or the media plan rather than the edit, say so early and we’ll point you to the right SkyLight team instead of quoting the wrong scope.
The production workflow behind a clean caption track
Straight up, here’s the order we run it. Transcription first, from the locked audio, so the timing anchors to the final edit and not a draft. Then translation and localization if a second language is in scope, handled by a native Arabic linguist rather than a machine pass. Then timing, where each cue is set to sit on screen long enough to read but not so long it lags the cut. Then QA, ideally a native pass for Arabic. Then delivery in the right shape: burned-in for social, soft SRT or VTT for web and long-form.
The sequence matters because captions built before picture lock get thrown out when the edit changes. We time against the finished cut, which is also why caption work lands in post rather than pre-production. A caption track is only as clean as the edit it’s timed to, so the two stay linked.
Before you wrap: if you want captions and the wider finish (grade, sound, deliverables) handled together, our color grading and post workflow runs alongside the caption pass on the same timeline.
FAQ
Do I need burned-in or soft subtitles for Instagram Reels?
Burned-in. Reels autoplay muted and crop aggressively, and uploaded caption files don’t display reliably in-feed, so baking the text into the frame with controlled placement is the safe choice.
Are Arabic subtitles more expensive than English?
Usually, yes. Arabic requires right-to-left layout, bidirectional handling for mixed English-Arabic lines, register choices and native QA, plus re-timing for text that runs 15-25% longer than English (reported). That’s more labour than an English pass.
What’s the difference between subtitles and localization?
Subtitles are the timed on-screen text of the spoken content. Localization goes further: it adapts idiom, formality and cultural phrasing so the Arabic reads like it was written natively, not translated word for word.
SRT or VTT, which format should I ask for?
SRT for general uploads like YouTube and LinkedIn; VTT (WebVTT) for website and HTML5 embed players where you want cue positioning and styling. Both must be saved as UTF-8 for Arabic to render correctly.
How long does a bilingual EN/AR caption pass take?
Reported bands run roughly 1-2 days for transcription and source timing, 2-3 days for translation and Arabic localization, and about 1 day for re-timing and integration, so a short-to-mid piece commonly lands within a working week end-to-end.
Can I just use auto-captions from CapCut or YouTube?
For a rough English internal clip, yes, after a cleanup pass. For paid campaigns, luxury brand films, client deliverables or anything in Arabic, auto-tools carry BiDi and register errors that read as sloppy to native viewers, so they’re not client-grade.
Why do so many videos need captions at all?
Reported silent-viewing rates of 80-85% on social feeds mean most viewers never hear the audio. On-screen text carries the message, and captioned video is also linked to longer watch-time and wider reach.
Do you produce the captions or just the video?
Both, as part of post-production. SL Media transcribes, times, translates, localizes and delivers the finished video with burned-in or soft captions in English, Arabic or both. Studio rental and media distribution sit with other SkyLight teams.