How do the papers measure quality degradation over rollout time (curves, drift deltas, judges), what do the numbers show, and where are the metrics unreliable?
Four measurement styles appear: (1) an imaging-quality-versus-time curve over 30 s (CausVid Fig. 8); (2) a first-5-s versus last-5-s imaging-quality delta, "ΔDriftQuality", on 30 s videos (Rolling Forcing); (3) whole-video VBench-Long dimensions at 50/75/100 s plus a Gemini-2.5-Pro "visual stability" score (Self-Forcing++); (4) VBench dimensions at 50/250 s with an explicit warning that copied clips fool consistency metrics (SVI). The authors of Self-Forcing++ and SVI both say standard image/consistency metrics can reward degraded or static video, so motion (dynamic degree) must be read with them.
Four measurement styles appear: (1) an imaging-quality-versus-time curve over 30 s (CausVid Fig. 8); (2) a first-5-s versus last-5-s imaging-quality delta, "ΔDriftQuality", on 30 s videos (Rolling Forcing); (3) whole-video VBench-Long dimensions at 50/75/100 s plus a Gemini-2.5-Pro "visual stability" score (Self-Forcing++); (4) VBench dimensions at 50/250 s with an explicit warning that copied clips fool consistency metrics (SVI). The authors of Self-Forcing++ and SVI both say standard image/consistency metrics can reward degraded or static video, so motion (dynamic degree) must be read with them.
Curator qualifications: All excerpts located; CausVid 30 s quality vs over-exposure at 50-100 s (different lengths/metrics) kept.
Evidence status: statements located in the cited sources by a correlated desk review (Claude); not executed; not independently reviewed.
Structured fields
{
"citations": [
{
"url": "https://arxiv.org/abs/2412.07772v4",
"note": "candidate evidence excerpts"
},
{
"url": "https://arxiv.org/abs/2509.25161v1",
"note": "candidate evidence excerpts"
},
{
"url": "https://arxiv.org/abs/2510.02283v1",
"note": "candidate evidence excerpts"
},
{
"url": "https://arxiv.org/abs/2510.09212v1",
"note": "candidate evidence excerpts"
}
]
}Evidence claims
- [sourced_fact] Four measurement styles appear: (1) an imaging-quality-versus-time curve over 30 s (CausVid Fig. 8); (2) a first-5-s versus last-5-s imaging-quality delta, "ΔDriftQuality", on 30 s videos (Rolling Forcing); (3) whole-video VBench-Long dimensions at 50/75/100 s plus a Gemini-2.5-Pro "visual stability" score (Self-Forcing++); (4) VBench dimensions at 50/250 s with an explicit warning that copied clips fool consistency metrics (SVI). The authors of Self-Forcing++ and SVI both say standard image/consistency metrics can reward degraded or static video, so motion (dynamic degree) must be read with them. · locator: candidate evidence excerpts source
Rights
| component | license | hosting | redistribution | modification | training | Evidence and notes |
|---|---|---|---|---|---|---|
| documentation | unknown | permitted | permitted | unknown | unknown | Applies to this original summary only, not to the cited sources. |