ByteDance Seedance Face Distortion — Pre-Generation Risk Reference
Technical Classification
Facial Landmark Regression Failure
Seedance produces strong static facial renders but landmark stability degrades across the temporal axis. Eye spacing shifts, jaw geometry reshapes, ear position drifts, and skin texture pixelates inconsistently. The failure is most severe in close-up shots and during head motion or facial expression changes. Seedance's support process is reachable through ByteDance's in-app support flow; technical-failure-coded requests with attached PDF audit reports route faster than generic complaints.
How to identify this failure
- ✕Eye spacing changes visibly across the clip
- ✕Jawline reshapes during head motion
- ✕Ear size or position drifts between frames
- ✕Skin texture quality inconsistent across frames
- ✕Facial expression interpolates through inconsistent intermediate forms
Real generation examples
Prompt used
"Close-up of a woman smiling, soft window light"
Failure observed @ 0:02
Left eye drifts inward by 0:02; jawline narrows by 0:04; expression mismatch between frame halves
Prompt used
"Man laughing at a dinner table, warm restaurant lighting"
Failure observed @ 0:01
Mouth shape inconsistent with prompted laugh; right ear morphs in size across 0:01-0:03
Documentation strength
If you need to escalate
VERY HIGH — Seedance support classifies facial landmark regression as a critical-tier failure. Documented requests with frame-level evidence approve quickly.
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 Facial Landmark Regression Failure 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 Facial Landmark Regression Failure 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 "Facial Landmark Regression Failure". 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
Will Seedance refund credits for face distortion?
Yes. Submit through ByteDance's in-app support flow. Use "Facial Landmark Regression Failure" terminology, cite the Generation ID, timestamp the worst-drift frame, and attach a PDF audit report. This is a high-precedent escalation category.
Why does Seedance distort faces across frames?
Seedance allocates per-frame compute aggressively to visual detail. The temporal-coherence module has limited budget for small high-frequency features (eye spacing, jaw geometry, ear placement), so these drift across the clip.
Which Seedance prompts highest-risk for face distortion?
Close-up portraits, prompts with facial expression changes, side profiles, multi-character scenes, and fast head motion. AVA's pre-flight scanner flags these patterns before generation.
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.
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.
Related failures across models
If you’re seeing this failure, you may also encounter these on other models:
Facial
Asymmetric eye placement, morphological drift across frames, non-Eucli…
Facial
Asymmetric eye placement, jaw drift between frames, non-Euclidean faci…
Facial
Sora output shows face morphing, identity inconsistency, or feature di…
Identity
Google Veo output shows facial feature distortion mid-clip or identity…
Identity
Face morphing, identity drift, asymmetric distortion, features melting…
Identity
Face morphing, identity drift, asymmetric distortion, features melting…
Pick a different tool for Seedance failures
Some prompt shapes will keep failing on Seedance. Routing those shots to a different vendor is the cheapest fix.