Last reviewed by a music video producer for production accuracy.
For a while, producing a music video meant one of two paths: spend thousands on a shoot, or spend weekends stitching together ffmpeg commands, Runway clips, and CapCut templates. Neither felt like a workflow — they felt like a part-time job. What changed my process was finding an AI music video generator that takes an audio file directly and handles the creative generation step in one go, producing a 9:16 vertical master sized for every distribution surface that matters today. This tutorial walks through the exact steps — from audio file prep to final export — using that pipeline from start to finish.
What You'll Need
Before running the workflow, confirm you have:
- An audio file in a supported format: MP3, M4A, WAV, AAC, OGG, or FLAC. AIFF is not accepted — export as WAV or FLAC from your DAW before uploading.
- Minimum track length of 60 seconds. Shorter samples will be rejected at upload.
- A file under 40 MB. Most stereo masters at 44.1 kHz / 16-bit WAV are well under this even at four minutes.
- An Echonos Pilot subscription ($30/month, 750 credits) — or 250 signup credits to run one test generation.
- A visual concept, even a rough one. More on prompt writing in Step 2.
Step 1 — Prep and Upload Your Audio
Go to the Echonos create page and upload your audio file. The uploader validates format and duration on drop — if either check fails, you'll see an inline error before any credits are used.
A few preparation tips that consistently improve generation results:
- Use a full mix, not a stem. The AI syncs visuals to the full audio energy profile; a dry vocal or lone guitar part produces weaker synchronisation than a mastered stereo bus.
- Export at 44.1 kHz, 16-bit WAV if your source is AIFF or a high-res format. The 40 MB cap is generous for standard resolution mixes.
- Trim silence from the top. The first few seconds of audio define the opening visual beat; dead air at the start produces a minimal opening that is hard to recover in Studio.
Step 2 — Write Your Visual Prompt
The prompt is your creative direction to the AI — it describes the mood, setting, colour palette, and visual language you want. The generator does not interpret song metadata or genre automatically; your prompt is the primary creative input.
Prompts that work well for music video generation typically include:
- A visual world: "rain-soaked neon Tokyo alley at night" anchors the setting more precisely than "dark and moody."
- A colour temperature: "warm golden hour" vs "cold blue tones" pulls the generation toward distinctly different palettes.
- A movement language: "slow-motion close-up of light refracting" vs "wide cinematic drone sweep" suggests how the camera should behave.
- A character reference (optional): upload a reference photo (up to 10 MB per image) for a consistent face or figure. Without a reference, the generation is purely scenic.
What to avoid: genre labels ("trap beat"), emotional abstractions ("sad"), and platform names ("TikTok video"). These give the model no useful visual anchor.
Step 3 — Run the Engine
Once the audio is uploaded and the prompt is set, confirm the generation settings and click Generate. Each full Engine generation costs 200 credits — a flat fee regardless of track length. A 90-second song and a 5-minute song cost the same.
What happens next:
- The AI analyses the audio for energy, tempo, and key-moment markers.
- Visual scenes are generated and synced to the audio waveform.
- The output is rendered as a 9:16 vertical master at 2K resolution — the native format for TikTok, Instagram Reels, YouTube Shorts, and Spotify Canvas.
Generation takes minutes, not hours. You'll receive a notification when the video is ready.
A note on aspect ratio: the 9:16 output is intentional, not a limitation. Every major short-form distribution surface — Canvas, Reels, TikTok, Shorts — is vertical-first. The vertical master IS the release asset for modern music distribution.
Step 4 — Review and Polish in Studio
Once the generation completes, the video opens in Studio — a scene-level editor where you can regenerate individual segments without re-running the full generation.
Studio fix costs are flat fees:
- Image scene regen: 10 credits per segment (the first 10 of a new subscription are free)
- Video segment regen: 50 credits per clip
The typical Studio pass for a 3-minute track takes 2–3 image regens and rarely needs a full clip regen. Budget 20–30 additional credits for a polish pass on top of the 200-credit Engine run.
According to the Spotify for Artists Canvas guide, Canvas clips perform best when the visual energy matches the song's most recognisable moments — use Studio to fine-tune scenes at the hook and chorus before export.
Step 5 — Export and Distribute
When the Studio pass is done, export the 9:16 master. The exported file is formatted for direct upload to:
- Spotify Canvas — upload through Spotify for Artists; Canvas requires a looping video in the 9:16 format Echonos outputs.
- TikTok — upload natively; the 9:16 master fills the TikTok screen without cropping or letterboxing.
- Instagram Reels — direct upload; 9:16 fills the Reels frame exactly.
- YouTube Shorts — upload as a Short; 9:16 is the required format.
For a deeper look at how this one-master workflow fits into a full release timeline, the music video without a camera guide walks through everything from mix day to Canvas upload.
Frequently Asked Questions
What audio formats does the AI music video generator accept?
The generator accepts MP3, M4A, WAV, AAC, OGG, and FLAC. AIFF is not supported — if your master is in AIFF, export as WAV or FLAC before uploading. The maximum file size is 40 MB, and the minimum track length is 60 seconds.
How many credits does it cost to generate a music video from an audio file?
Each full Engine generation costs 200 credits regardless of track length. Studio polishing adds flat fees: 10 credits per image scene regen (first 10 free on a new subscription) and 50 credits per video clip regen. A Pilot Plan subscription (750 credits at $30/month) covers roughly three full Engine generations with headroom for Studio fixes.
Can I generate a horizontal (16:9) music video from an audio file?
Not currently. The generator outputs 9:16 vertical only, sized for TikTok, Instagram Reels, YouTube Shorts, and Spotify Canvas. Horizontal output is on the roadmap. For YouTube main-feed 16:9 uploads, a separate horizontal-output step outside Echonos is needed today.
How long does AI music video generation take?
Generation takes minutes, not hours. Most tracks complete well within a work session. The Engine analyses audio, syncs scenes, and renders the 9:16 master without any real-time preview delay.
Is there a free trial or free plan?
Echonos does not have a free subscription tier. New accounts receive 250 signup credits, which cover one full Engine generation (200 credits) with roughly 50 credits of headroom for a Studio fix. After the signup allocation, the live subscription is the Pilot Plan at $30/month with 750 credits.
What makes a good visual prompt for AI music video generation from audio?
Prompts work best when they describe a specific visual world — setting, colour temperature, and camera movement. Avoid genre labels and emotional abstractions. "Neon-lit rain on a Tokyo street, close-up droplets, cold blue tones" outperforms "sad alternative music video" by a wide margin.
Final Thought
The core workflow — upload, prompt, generate, polish, export — takes less than an afternoon the first time and gets faster with each release. The 9:16 output is not a compromise; it is the correct format for where audiences watch music today. If the only thing holding up your release visuals is the production step, this pipeline removes it.
About the Author
This workflow was developed and tested across a series of independent short-form music releases. The author has produced and directed music videos for independent artists, with a focus on vertical-first distribution pipelines and AI-assisted production for budget-conscious releases. Opinions are based on direct production experience.
Disclosure: This article contains contextual links to Echonos, an AI music video tool. The workflow described is based on direct use of the product. No payment was received for this article.
