CRITICALFailure Reference

Veo Face Distortion — Refund Guide

Technical Classification

Identity Embedding Drift

Identity Embedding Drift on Veo occurs because video diffusion models encode face structure as a regional likelihood prior rather than as a persistent identity embedding. Across frames, the per-frame face-likelihood prior re-rolls and the rendered face drifts in eye spacing, jaw shape, or skin texture. In multi-shot prompts the drift compounds across cuts and the character becomes unrecognisable. This is a critical defect for any branded or character-consistent use case.

How to identify this failure

✕Eye spacing or jaw shape shifts between frames of a continuous shot
✕Skin texture re-rolls — pore pattern changes per frame
✕Character identity fully changes between two cuts of the same person
✕Facial expression shifts independent of prompt direction

Real generation examples

Prompt used

"Close-up of a woman smiling, then turning to the side"

Failure observed @ 0:02

Jaw line and eye spacing visibly change during the turn — feels like two different people

Documentation strength

If you need to escalate

HIGH — Google Veo support refunds face-drift tickets on paid output when timestamps are provided.

AVA is a pre-purchase prevention tool, not a post-purchase recovery tool. Platforms generally do not guarantee credit refunds for output-quality failures; goodwill credits are at each platform's discretion. The strength rating reflects how well-formed your support ticket can be, not a promised outcome.

Prevention + documentation steps

01
Score your prompt before you generate
Run your prompt through AVA's pre-flight scoring against the Identity Embedding Drift pattern. Green light = generate. Yellow/red = rewrite using the suggested fix before you commit credits.
02
Capture Generation ID + timestamp if it failed anyway
Find the Generation ID in the URL or share link. Note the exact time when the Identity Embedding Drift first appears (e.g. "failure first visible at 1.2s"). Timestamped evidence is significantly stronger than a general complaint.
03
Use the correct technical term in your support ticket
Describe this failure as "Identity Embedding Drift". This term maps to a recognised internal workflow in the support system and routes the ticket to the right team.
04
Submit via the correct support channel
Runway has no direct email intake. Pro+ plan: open the in-app AI Assistant (help widget bottom-right of app.runwayml.com), describe the failure with the technical term, attach evidence. Free/Standard plan: human support isn't available — your channel is Discord #community-help with @On Call - Moderators.

Frequently asked questions

Does Google refund Veo credits for face distortion?

Yes — Veo treats identity drift on paid output as a defect. Cite "Identity Embedding Drift" with a still frame showing the change.

Why does Veo distort faces across frames?

Video diffusion has no persistent identity embedding — face structure is a per-frame likelihood prior that re-rolls. Across long clips or cuts, the rolls drift and identity changes.

How do I avoid Veo face drift?

Keep face on screen for shorter durations. Avoid cuts within a single generation. Use reference images where supported.

Score your prompt

Score your prompt against this failure mode in 30 seconds

Paste your prompt and the platform you intend to use. AVA returns a red/yellow/green score against this specific failure mode plus a concrete rewrite if the risk is high.

Install free All 105 failure modes

AVA Pro · founders' round

$50 for 6 months of unlimited scoring across all failure modes + personal failure-history dashboard. Locks in $13/mo grandfathered after.

Claim $50 founders