How to Create Custom YouTube B-Roll with Midjourney
The High Cost of Generic Stock Footage
Last winter, I spent three hours scrolling through stock video platforms trying to find a four-second clip of a neon-lit Seoul alleyway in the rain that didn’t look like a generic corporate travel ad. When you run multiple YouTube channels as a solo creator, stock footage quickly becomes a bottleneck. You either pay hundreds of dollars a month for high-end stock subscriptions or end up using the same overused clips as five other channels in your niche.
I started experimenting with Midjourney to generate custom visual assets for my video edits. At first, the results were static and obviously AI-generated. But after refining my prompting workflow and combining Midjourney images with lightweight video motion tools, I built a reliable system for producing unique, high-quality YouTube B-roll on demand. Here is the exact workflow I use across my channels today.
Step 1: Prompting Midjourney for Cinematic Framing
The biggest mistake creators make when using Midjourney for YouTube B-roll is writing prompts designed for static artwork rather than cinematic footage. Midjourney defaults to stylized, centered subjects unless you explicitly instruct it to mimic real camera lenses and cinematography.
To generate images that feel like video stills, you need to include specific aspect ratios, camera angles, and lighting parameters in every prompt.
Essential Parameters for Video Assets
Always append the correct aspect ratio flag to your prompts depending on the platform:
- For standard YouTube (16:9): Add
--ar 16:9to the end of your prompt. - For YouTube Shorts (9:16): Add
--ar 9:16to the end of your prompt.
Beyond aspect ratios, camera placement dictates how convincing your shot will look once animated. Rather than asking for a simple subject, specify camera framing like extreme wide angle shot, over-the-shoulder perspective, or macro close-up with shallow depth of field.
A Practical B-Roll Prompt Formula
I use a structured template for generating consistent background visuals:
[Subject or environment] + [Shot type / camera angle] + [Lighting style] + [Color palette] + --ar 16:9 --v 6.1
For example, if I need B-roll depicting a software developer working late, a bad prompt is programmer at laptop working late. A production-ready B-roll prompt looks like this:
Over-the-shoulder shot of a modern laptop screen displaying code, dark moody home office, soft blue light from the monitor, ambient warm desk lamp blur in background, cinematic 35mm film photo, shallow depth of field --ar 16:9 --v 6.1
Step 2: Maintaining Visual Consistency Across Cuts
A major giveaway of amateur video editing is jumping between wildly different visual styles in the same scene. If shot A looks like an anime illustration and shot B looks like a raw photograph, your viewer will immediately disengage.
Using Style References (–sref)
Midjourney features a Style Reference parameter that lets you maintain visual continuity across multiple generations. Once you generate a base shot that hits the aesthetic you want, copy its image URL and append it to subsequent prompts using the --sref [URL] parameter.
When I generate a sequence of shots for a single video segment, I create one primary establishing shot, grab its URL, and use that as the style anchor for all secondary close-ups and reaction shots in that segment.
Locking Your Color Palette
In addition to style references, keep your lighting and color descriptors identical across prompts within the same video topic. If your first shot uses cool morning light, muted color grading, carry those exact phrases through all prompts for that sequence.
Step 3: Animating Static Midjourney Images into Motion B-Roll
Static images on YouTube feel flat if left on screen for more than two seconds. To turn Midjourney outputs into genuine B-roll footage, you need to introduce movement.
Method A: Simple Pan and Zoom Keyframing (No Extra Tools Needed)
You do not always need AI video generators to animate B-roll. The simplest and most reliable method is applying a classic Ken Burns effect inside your video editing software, such as CapCut or Premiere Pro.
- Import your 16:9 Midjourney image into your editing timeline.
- Set a keyframe at the start of the clip with a scale of 100%.
- Set a keyframe at the end of the clip (3 to 5 seconds later) with a scale of 108% and a slight horizontal pan.
- Apply an easy-in/easy-out curve so the motion feels natural.
Because Midjourney outputs high-resolution images, you can scale them up significantly without losing sharpness. This subtle motion keeps the viewer’s eye engaged without creating visual artifacts.
Method B: AI Image-to-Video Generation
When you need real object movement—such as drifting smoke, moving vehicle lights, or natural atmospheric motion—pass your Midjourney generation into an image-to-video tool. CapCut has built-in image animation features, while standalone tools like Runway or Luma offer advanced motion controls.
When animating AI images into video, less is more. Set motion parameters low. High motion settings often cause backgrounds to warp, textures to melt, or human faces to distort. Subtle camera pans, gentle water reflections, and ambient atmospheric drift produce the cleanest results.
Comparing B-Roll Creation Methods
To help you decide when to use Midjourney over traditional methods, here is how the primary approaches compare for a solo YouTube operation:
| Method | Speed & Cost | Visual Uniqueness |
|---|---|---|
| Midjourney + Motion | Fast / Low ongoing cost | 100% Unique to your prompt |
| Stock Video Libraries | Fast / Subscription fees | Low (Overused across channels) |
| Live Filming | Slow / Equipment costs | High (Limited by location/gear) |
Integrating Midjourney into Your Editing Pipeline
To avoid spending hours generating images during the editing phase, batch your visual asset creation before you open your video editor.
- Script Formatting: When drafting your script in Notion or ChatGPT, mark visual cues in a separate column alongside your audio narration.
- Batch Generation: Open Midjourney and generate 3 to 4 prompt variations for each visual cue. Upscale the best option for each segment.
- Organized Export: Save your assets into dedicated topic folders named after script sections (e.g.,
01_Intro_Background,02_Tech_Closeup). - Editing Assembly: Drag the folder directly into CapCut or your preferred editor, apply your movement presets, and trim clips to match your voiceover rhythm.
My Take: Honest Strengths and Limitations
Midjourney has transformed how I produce visuals for my automated content channels, but it is not a silver bullet for every video editing scenario.
Midjourney excels at establishing shots, atmospheric backgrounds, conceptual illustrations, historical scenes, and abstract technology visuals. It allows you to create imagery that would otherwise require a full production crew or an expensive location shoot.
However, Midjourney struggles with specific technical details. It is terrible for generating accurate software user interfaces, readable text overlays, or consistent multi-step hand movements (like typing on a keyboard or drawing on paper). Trying to force Midjourney to generate a precise UI screenshot will waste your time. For real-world software, product demos, or exact step-by-step physical tasks, traditional screen recording or real footage remains far superior.
Regarding costs, Midjourney operates on a paid subscription model (typically starting in the $10 to $30/month range depending on fast-gpu usage tiers). Motion generation tools or editing suites like CapCut offer free tiers alongside paid plans for advanced features. Pricing schedules change frequently, so always check the official pricing pages for Midjourney, Runway, and CapCut before building your tech stack.
FAQ
How do I get a 16:9 aspect ratio in Midjourney for YouTube?
Add --ar 16:9 to the very end of your prompt text string. If you are creating content for YouTube Shorts, Reels, or TikTok, use --ar 9:16 instead.
Can I animate Midjourney images into video B-roll for free?
Yes. You can import static Midjourney images into free video editors like CapCut and use keyframing to create manual zooms and pans. Additionally, several video editing applications offer limited free trial credits for AI image-to-video motion generation.
Will YouTube penalize channels for using AI-generated B-roll?
YouTube requires creators to disclose when content is altered or synthetically generated in ways that could mislead viewers about real events. However, using AI-generated background imagery, conceptual illustrations, or stylistic B-roll as visual support for an original voiceover or presentation generally falls under standard content creation techniques. Always review YouTube’s latest synthetic content guidelines to ensure full compliance.
