DeepMist AI

The demo reel  /  MISK breakdown

Production Breakdown

An Arabic ad for a scent you cannot film

This page is a full production breakdown of a short vertical ad for MISK, a fictional oud perfume we invented for our own portfolio. Nobody hired us for it. In the Gulf, perfume and oud is among the biggest online categories, and the camera has a problem: it cannot carry a scent. So the ad sells the effect instead.

MISK is a brand we invented. Self-initiated demo, not client work.

The finished ad. 12 seconds, 9:16, sound on. Spoken in Modern Standard Arabic; the words are written out below in Arabic and in English. Every frame AI-generated. The “portfolio demo” watermark is ours — client deliveries ship clean.

The brief we wrote ourselves

Product
A Gulf oud perfume that exists only as this demo.
Audience
Gulf women shopping for perfume and oud online, on their phones.
Constraint
No filmed footage, a single continuous shot, runtime locked to the voiceover.
Format
Mirror-selfie · talking head · spoken Arabic
Niche
Perfume & Oud

The opening line, as it is spoken

رشّة واحدة من مسك، وتبقى معي طوال اليوم.

In English: “One spray of MISK, and it stays with me all day.”

The whole spoken line

رشّة واحدة من مسك، وتبقى معي طوال اليوم. الجميع يسألني: ما هذا العطر؟ جرّبوه، لن تندموا.

One spray of MISK, and it stays with me all day. Everyone asks me: what is that perfume? Try it, you won't regret it.

The English here is a translation for this page. It is not spoken in the ad, and it is not burned into the video.

The opening line makes a claim you can hold in your head: a single spray lasts the day. That matters because video cannot carry a scent. So instead of describing the perfume, the line describes its effect, and the follow-up adds the social proof: people ask her what it is.

Why this format. Perfume sells on personal recommendation, not on spectacle. A mirror selfie with a face talking straight to camera looks like something a friend would send, not a brand announcement. It keeps the bottle in a human hand, next to a human face.

How it was built

  1. 01

    Script first

    The spoken line is written in Modern Standard Arabic before any image exists, and it is measured by ear rather than by word count. A script that looks short on the page can run long once it is read aloud, so it gets cut back until it sits inside the time available.

  2. 02

    Voice before picture

    The Arabic voiceover is generated with a text-to-speech model and approved as a standalone file before a single frame is drawn. This is the inversion from our English ads, where the video model generates the audio inside the same pass that generates the picture.

  3. 03

    Hero still

    One vertical image, generated with Nano Banana Pro. Framed like a phone photo rather than a studio product shot: the presenter and the product in hand, the brand name legible, and light with a plausible source in the room.

  4. 04

    Motion and lip-sync

    The still goes to Seedance 2.0 as frame one, with the approved audio file alongside it as a reference, so the mouth moves to the actual Arabic speech instead of a dub laid over it afterwards. The audio leads the picture here rather than the other way round, which is what the lip movement is following.

  5. 05

    Check before keep

    Nothing survives because it looked fine as a thumbnail. We read the stream data with ffprobe, build a contact sheet of frames with ffmpeg, and compare the spoken audio against the written script. Takes that drift from the brief get regenerated, not patched.

  6. 06

    Finish

    The approved take is cut and packaged with ffmpeg for wherever it will run. Vertical for Reels and TikTok here; the same take can be reframed for feed or landscape if the brief asks.

Format
9:16 vertical, 1280×720 source
Runtime
12.05 seconds, audio included
Stills
Nano Banana Pro
Motion & lip-sync
Seedance 2.0, image-to-video
Voice
Inworld, Arabic text-to-speech
Ad language
Modern Standard Arabic
Assembly & QA
ffmpeg / ffprobe
On location
Nobody. No crew, no samples shipped

What to watch for

Lip timing

Watch her mouth: the lips move to the real Arabic words, because the voice was made before the picture, not dubbed on after.

Phone-photo framing

The still was framed like a phone photo, not a studio shot, so the mirror and window light read as everyday, not staged.

Bottle handling

She keeps the bottle near her face throughout, the way people actually show a perfume, so the product stays inside the conversation.

Two things about Arabic we will not smooth over

The logo on the product is in Latin letters, not Arabic ones. The reason is technical and known: image models do not render connected Arabic script reliably, and it comes out malformed. We chose a clean Latin wordmark over broken Arabic type. If your brand's Arabic logotype has to appear on the product in frame, say so at the brief stage, because it changes how the shot is built.

The voice is Modern Standard Arabic, and it is generated, not a hired voice actor. That is the right register for a pan-Gulf audience and the wrong one if you specifically need spoken Khaleeji dialect or one particular human voice. It is the single thing on this page you cannot judge for yourself if you do not speak Arabic, so we would rather write it down than let the video imply otherwise.

What this page is not

Every frame is AI-generated. The brand is invented. The ad never ran as paid media, so no performance numbers are claimed. The voice is AI-generated Modern Standard Arabic, and buyers needing spoken Gulf dialect or a specific voice should not order this. A real product would get real close-ups.

We have no published reviews yet, so we are not going to borrow credibility we haven't earned. No client names, no engagement numbers, no invented case studies. What we can show you is the work and the method, in full, and let you judge whether it clears your bar.

Want an opening line like this for your product?

Get 3 custom hooks — free Or skip ahead and brief us on Fiverr to have the whole ad built — script, presenter and edit.