AI Short Film 'Reflections Lie Still': How to Recreate 1980s Dark Fantasy Film Aesthetics with Midjourney

How one creator used Midjourney to systematically build a 1980s dark fantasy aesthetic for the AI short film 'Reflections Lie Still.'
The making of 'Reflections Lie Still' charts a mature path for AI filmmaking: establish a complete visual language in Midjourney before generating a single shot — anchored in 1980s practical-effects fantasy and VHS grain, inspired by the dark grandeur of Elden Ring, and deliberately imperfect to escape the AI 'digital look.' The creator's shared prompts reveal four cinema-grade techniques: emphasizing spatial scale, describing light behavior over objects, creating spatial instability, and closing with an emotional target. The real lesson is methodological — clear aesthetic direction, systematic concept design, refined prompts, and a complete narrative framework are what truly differentiate AI filmmaking.
In an era where AI filmmaking is rapidly maturing, the ability to give generated visuals a unified and distinctive aesthetic language has become the defining line between a "tech demo" and a genuine work of art. One creator answered this challenge with the short film Reflections Lie Still — using Midjourney to systematically build a complete 1980s dark fantasy aesthetic, then constructing every shot in the film around it.

From Concept Art to Visual Language: The Pre-Production Art Direction Workflow in AI Filmmaking
Before shooting a single frame, the creator from Prowler Productions generated a series of concept images in Midjourney — not to use directly, but to "lock in the visual language." This is a critical step in traditional film production: establishing the tone, environment style, and lighting direction of an entire work through concept design, long before actual filming or modeling begins.
He explicitly set his aesthetic references as "the atmosphere of old-school practical effects fantasy films" and "the grainy, VHS-era image quality." This approach is well worth borrowing for AI creators: rather than letting each frame fight for its own identity, establish a clear aesthetic reference system first, so that all generated content converges around the same core style. This also directly addresses one of the most common pain points in AI content generation — "style drift."
Style Drift is a common frustration in AI image generation. Because each generation is an independent random sampling process, even with similar prompts, consecutive outputs can vary significantly in color tone, compositional language, and detail density — resulting in visual inconsistency within a single film. Traditional film production uses "Bible" documents and moodboards to anchor style. AI creators need a similar mechanism: generate a batch of highly consistent reference images to establish a "style seed," then use Image Prompts and style references (
--sref) in subsequent generations to pull new frames toward that seed. This artificially creates overall visual cohesion even without the physical constraints of a real film set.
A Dark Fantasy Story About Mirrors
Reflections Lie Still follows a cloaked figure known as the "Mirror Warden," journeying through a mysterious world filled with ancient mirrors. Each mirror reflects not just space, but traps creatures within its surface — and the Warden's journey takes it through the eerie architecture that surrounds them.
The film is narrated in a mythic, poetic first-person voice from the Mirror Warden's perspective, guiding viewers through fog-shrouded landscapes, crumbling towers, and unsettling scenes where nothing is quite as it appears. Shattered mirror halls, cliff ruins veiled in mist, shadow-drenched forests — every space is imbued with a sense of weight, mystery, and deeply unsettling grandeur.
The creator openly acknowledged that the atmosphere of these environments drew heavily from the world-building of Elden Ring and Dark Souls. Both games are renowned for their oppressive yet magnificent environmental design, making them key reference points for contemporary dark fantasy aesthetics — and a clear sign that game visuals are increasingly influencing AI filmmaking in return.
Crafting That '80s Feel: Stripping AI Visuals of the "Digital Look"
Worth particular attention is the creator's deliberate pursuit of period authenticity. He wanted the entire film to feel like "a dark fantasy movie lost in the 1980s," and emphasized three aesthetic pillars to achieve this:
- Practical effects texture: Simulating the physical quality of models and props from the film era, rather than the clean, smooth look of the digital age.
- Immersive lighting: Using light to create atmosphere rather than to impress.
- Heavy ambient presence: A combined environment of fog, grain, and shadow.
This counter-intuitive pursuit of "imperfection" is genuinely instructive. When the default tendency of AI-generated imagery leans toward high resolution, sharpness, and saturated color, deliberately introducing grain, haze, and period-appropriate filters can actually liberate the work from feeling "AI-generated" — granting it stronger artistic identity and emotional weight.
The "Digital Look" refers to the characteristic over-sharpness, dense high-frequency detail, and overly "correct" lighting common in AI-generated images. This stems from diffusion models being trained predominantly on modern high-resolution photography, biasing default outputs toward technical perfection. In contrast, 1980s film imagery carries the marks of its physical limitations: grain from low-sensitivity stocks like Kodak 5247, soft bokeh caused by spherical aberration in period lenses, slight imperfections in handcrafted model props, and the distinctive oval bokeh and lens flares produced by anamorphic lenses. Incorporating descriptors like
35mm film grain,anamorphic lens flare, andpractical effects textureinto your prompts — combined with film grain overlays in post-production — can effectively steer AI output toward a period aesthetic, visually distinguishing your work from the ocean of homogeneous high-definition AI content.
Prompt Breakdown: How to Write Cinema-Grade Scene Descriptions
The creator publicly shared the complete prompt for the key concept frame — the "Infinite Mirror Hall" — and it's a masterclass worth studying. It depicts a circular megastructure: composed of countless mirror-smooth columns through which light refracts endlessly; a grand hall of dark stone where reflective panels cover every wall, floor, and ceiling, scattering images into dizzying, seemingly infinite patterns.
Breaking down this prompt reveals several key techniques:
1. Establish Spatial Scale
Words like "vast," "towering," and "immense" are used repeatedly to hammer home a sense of scale, giving the image an inherent, oppressive grandeur.
2. Describe the Behavior of Light, Not the Objects Themselves
The prompt spends considerable space on how light "refracts endlessly," "scatters," and "bends in unpredictable directions." Describing the behavior of light rather than objects directly tends to produce far more dramatically compelling imagery.
3. Create Spatial Instability
"Suspended shards of broken glass," "jagged stone platforms rising and falling irregularly," "precarious pathways" — these details collectively build an "other-worldly instability" that serves the film's unsettling atmosphere.
4. Close With an Emotional Target
The prompt ends by summarizing the overall effect as "a hypnotic interweaving of solid forms and their reflections" — essentially giving the AI a clear emotional goal, rather than just piling on objects.
The core difference between cinema-grade prompt writing and ordinary image generation prompts lies in whether the writer adopts a cinematographer's perspective. Ordinary prompts tend to enumerate "what is in the frame," while cinema-grade prompts must simultaneously address three layers: Subject (what is being captured), Lighting & Space Behavior (the physical behavior of light and space), and Emotional Tone (the emotional target). The analogy to storyboarding is apt: a director doesn't just tell the DP "there's a mirror here" — they say "light comes in at 45 degrees from the left, creating a scorching quality on the mirror surface, making the whole space feel like it's delivering a verdict." Translating this directorial mindset into prompts means helping the AI understand why the image exists, not just what exists in it.
Methodological Takeaways for AI Filmmakers
The complete short film Reflections Lie Still is now available on the Prowler Productions YouTube channel. As an AI filmmaking experiment completed by an individual creator, it demonstrates the current maturity of the workflow: use a text-to-image tool to lock in the aesthetic first, then build a complete narrative film around it.
The greatest value of this case study lies not in the technology itself, but in the creative methodology: a clearly defined aesthetic direction + systematic concept design + refined prompt engineering + a complete narrative framework. As more and more creators gain access to AI tools, true differentiation will come back to taste, storytelling, and stylistic consistency. The tools lower the barrier to entry — but artistic judgment remains rare.
Related articles

Vercel AI SDK Releases Vue 3.0.282 Patch Update
Vercel AI SDK releases @ai-sdk/vue@3.0.282 patch update, syncing with core package ai@6.0.282. Learn about the changes, release cadence, and upgrade recommendations.

Vercel AI SDK Sandbox Component Receives Patch Update
Vercel AI SDK releases sandbox-vercel@1.0.109 patch update, syncing the harness dependency to the same version. A look at this maintenance release and what it means for AI app developers.

Vercel AI SDK Vue 4.0.99 Released: Dependency Update Overview
The @ai-sdk/vue 4.0.99 patch release syncs the underlying ai@7.0.99 dependency. Learn what this means for Vue developers building AI apps with Vercel AI SDK.