Behind the scenes · Lyric video production
AI Slop or True Craftsmanship? How a Lyric Video Truly Comes to Life
We are living in an era where AI is spoken of as if it were a magic wand. Suno spits out full songs in seconds, image generators conjure visually impressive artwork in an instant, and video generators create moving pictures from a single prompt.
When AI is utilized in a lyric video, the knee-jerk reaction these days is unfortunately all too predictable: “Oh, look, AI slop! Someone just pressed a button and let the machine do everything.”
People are, in a way, right - but at the same time, fundamentally wrong. This stems mostly from a lack of understanding.
Yes, the internet is flooded with lazy, automatically generated trash pushed into marketing and social media without any vision or skill, and in many ways, that is wrong. But that world is light-years away from what I, and many other professionals, do.
This same general misconception often pops up when talking about using AI in software development. Every single dev shop uses AI today, resulting in better code delivered faster and more efficiently than ever before. But to achieve that, it requires strict validation, precise processes, and refined workflows behind the scenes.
Yet many people assume AI coding is inherently flawed and risky, simply because the internet is also full of non-professionals "vibe-coding" apps that naturally lack professional expertise, validation, and rigorous processes. But that has nothing to do with professional production.
Raw Material Is Just Raw Material
I am not a videographer; I am a video editor.
At the end of the day, I don’t care where the raw footage comes from. It could be shot by a band on a smartphone, filmed by a full camera crew, sourced from free or paid stock libraries, built in 3D software, assembled from 2D images, hand-painted - or generated as pieces using AI.
My role is not to press a "generate" button. My role is to edit and refine the material, weave it together, and craft a cohesive whole.
Budget constraints are common in lyric video production; we are usually talking about projects under 1,000 euros. That budget doesn't leave room to fly a camera crew to Norway for fjord shots, nor does it allow for weeks of 3D rendering. A single shoot day easily costs €1500 for just one camera operator, excluding travel expenses, gear, props, or editing - which takes considerably more time than a single day on set.
This is where AI has solved the fundamental issue of sourcing raw assets:
- Stock Footage: High-quality stock is expensive, and free libraries are severely limited. That specific vision you need is rarely waiting for you in a stock footage library, paid or free.
- 3D Animation: I can build complex scenes in Blender, but it takes an immense amount of time relative to the budget. The entire video budget would easily be consumed by a single 5-second 3D scene.
AI acts as a modern stock footage bank for me, from where I can search fitting pieces to my video. But while many assume the work is finished at that point, on my desk, it is only just beginning.
The Work Hours Are Still There
Believe it or not, working on a single lyric video still takes tens of hours of active work time. It involves a massive amount of 3D tracking, masking, rotoscoping, manual animation, building custom animation templates, programming camera animations, tweaking and matching VFX effects, and of course, continuous design and experimentation. Behind it all lies years of experience, professional software, expertise in utilizing these tools, and a library of plugins, textures, and VFX assets built up over time.
This is the classic curse of the VFX artist: the viewer won't consciously notice most of the work you've put in - but they would immediately notice if it were missing.
AI has not shortened the overall process or eliminated manual labor; it has simply shifted its nature, particularly regarding asset acquisition.
Generating raw footage with AI is often just step one, followed by a substantial technical and artistic correction process before the actual editing even begins:
- Quality Enhancement: AI assets are often low-resolution and have poor frame rates (FPS). Both need to be processed through specialized tools just to make the footage usable.
- Intensive Manual Labor: Because AI footage cannot be planned or shot specifically for the edit (no tracking markers, no controlled camera angles or framing), it demands a massive amount of manual tracking, rotoscoping, and masking.
- Fixing Artifacts: AI outputs almost always contain anatomical flaws, logic inconsistencies, and visual glitches that must be manually patched and repaired.
I work in the Adobe ecosystem, meaning I rely on heavy-duty software. Every single video uses at least After Effects and Photoshop, and if live-action footage is involved, Premiere Pro handles the cut. Additionally, Blender runs in the background whenever genuine 3D environments or objects are needed.
AI Is Here to Stay - And It’s Already Built into Professional Software
It is also worth understanding that AI isn't just a collection of standalone image or video generators. It has been deeply integrated into industry-standard creative suites.
Modern Photoshop, After Effects, and countless other professional applications use AI in many of their core functions: background removal, masking, generative fill, tracking, color grading, and more. This isn't taking shortcuts - it has brought a massive leap in quality, precision, and workflow efficiency to these tools.
The Final Result Is What Matters
When the track plays and the screen lights up with a dark, polished lyric video cut to the rhythm of the music, the viewer experiences the song visually. That experience shouldn't be defined by where the raw pixels originated, but by how they are edited, color-graded, and paced to tell a story.
AI won't make a finished video for anyone. It is simply one tool among many; a tool that allows bands with smaller budgets to bring unique visual worlds to life. But wielding that tool still demands tens of hours of labor, an eye for editing, and solid craftsmanship.
The machine generates raw material. The editor creates the video.
P.S. There are already several "AI lyric video generators" on the market. Anyone is free to try them out to see what kind of output they produce and how limited their capabilities actually are.
Looking for a lyric video that goes beyond AI generation?
A strong lyric video combines visual direction, precise editing, typography, and a clear understanding of the song. If you are preparing a release and want a video built around your music, take a look at my lyric video production services or get in touch to discuss your project.
