How to make a macro assembly video of a product with tactile sound

Last updated 2026-08-16 · 6 min read

What is the short version?

A close-up assembly video is one of the most watched product formats there is, and filming one properly means a macro lens, a rig, controlled light and the actual object. Generating one means describing it precisely and then judging the sound as carefully as the picture.

1. Go to Vizard Agent and describe the object and its surface.
2. Say what is assembled and in what order.
3. Say the pacing and the length.

What do you need before you start?

A precise description. There is no footage here, so the specificity of the brief is the whole input — the material, the finish, the surface it sits on and the tools involved all change what gets generated and what it sounds like.

• The object. Material, colour, finish, scale. "Matte black" and "gloss black" are different videos.
• The surface. A wooden bench, a white studio table, a dark cloth. It sets the whole look.
• The sequence. Which parts go on in which order, and with what tools.
• The pacing. Fast-cut, or slow and deliberate. This decides the shot count.
• The length. Around fifteen seconds is a working size for a feed clip.

What do you type into Vizard Agent?

Write it like a shot description with materials named. Vizard Agent generates from the specifics, so the words that earn their place are the physical ones — the finish, the tool, the surface — rather than adjectives about how satisfying it should feel.

Prompt

Variants worth knowing:

• Slow rather than fast-cut. A held macro shot suits some products far better than rapid cutting.
• Match an existing video's pacing. Reference one and Vizard Agent works to that rhythm.
• Silent version. Useful if the video will run under a voiceover or a platform's own audio.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real macro assembly clip. It spends a shot proving the look before spending four more, and it treats the sound design as a second pass rather than an afterthought.

1. Checks the workspace and the available tools.
2. Searches for existing footage of the subject before generating anything.
3. Checks the generation options and their pricing, along with the sound and image tools.
4. Searches the stock library in case the shot exists already.
5. Generates the first shot alone — the engine assembly with precision tweezers.
6. Extracts preview frames from that shot and looks at them before going further.
7. Generates shots two to five concurrently, then tiles their frames and reviews each in turn.
8. Generates custom sound effects for the tactile actions and checks every asset's duration.
9. Assembles the cut with synchronised sound, inspects the timeline frames and the waveform, and re-renders with a crisper mix.

Step 5 is the discipline. One shot costs a fraction of five, and inspecting it first is how a wrong material, a wrong scale or a wrong light gets caught before it is repeated four more times.

What does the result look like?

From the run this page is written from, probed on the delivered file: 1920x1080, H.264, 24fps, 13.0 seconds, AAC audio. Wide, thirteen seconds, five generated macro shots cut together with sound designed per action and a mix rebuilt after the first render was reviewed.

Vizard Agent then ran an analysis over the finished file to check the pacing, the continuity between shots and the synchronisation of the sound with the picture.

Thirteen seconds across five shots is under three seconds each. That is the range where generated macro footage holds together, and it is also the rhythm this format wants — each action registers and none of them lingers.

Timing was not measured separately for this kind of job. The closest measured work runs a median of 28 to 38 minutes end to end, with the middle half spread considerably wider. Across all projects the median cost by tier is Flash 47, Pro 55, Max 242, Ultra 263 credits.

When does this not work well?

Macro generation is convincing on materials and textures and unreliable on mechanics and hands. Vizard Agent inspects the first shot before committing to the rest and reviews every one afterwards, and there are things this format simply cannot show truthfully no matter how carefully it is briefed.

• It is not your product. A generated model resembles the description, not the item you sell. For a real product, film it.
• Assembly logic is invented. The parts go together plausibly rather than correctly. Do not use this as an instruction video.
• Hands and tools are the weak point. Tweezers holding a small part is exactly where generated video strains.
• Text and logos come out wrong. Anything with branding legible in frame will not survive.
• The sound is designed. It is convincing and it was not recorded from the object, which matters if your audience believes otherwise.
• Five shots is the practical limit. Each one is a separate generation, so a longer sequence costs proportionally more.

How do you fix a result that came back wrong?

Name the shot or the sound. Vizard Agent keeps every generated shot and each sound layer separately, so regenerating one angle or sharpening the foley is a change to one piece rather than a rebuild of the whole cut.

• "Shot three is the wrong scale." Regenerated against the same description.
• "The sound is too soft." A crisper mix from the same layers.
• "Slow the cutting down." A re-pace across the existing shots.

How does Vizard Agent compare to doing it yourself?

By hand this is a macro lens, a stop-motion rig, controlled lighting, the actual object, and a foley pass — a day of setup for thirteen seconds, which is why this format is dominated by studios rather than by the people who make the products.

Common questions

Can it show my actual product? Not exactly. Vizard Agent generates from your description; for the real item, filming it is the honest route.

Can I use it as an assembly guide? No. The sequence is plausible rather than accurate, and Vizard Agent has no knowledge of how your product actually goes together.

Does it generate the sound? Yes. Vizard Agent produces a sound effect per tactile action and syncs it to the cut.

How long should it be? Ten to fifteen seconds. Every extra shot is another generation, so length has a real cost.

Can I ask for slower pacing? Yes, and it suits many products better. Say so and Vizard Agent holds the shots longer.

Can it match a reference video's rhythm? Yes. Point Vizard Agent at one and it works to that pacing rather than inventing its own.

Can it use my own macro footage? Yes, and it is better. Vizard Agent will cut your material and design the sound over it.

Should the video have music? Usually not. This format lives on the foley, and Vizard Agent builds the layers rather than scoring over them.