The export burns subtitle text and its black bar into the video pixels instead of keeping them as a separate editable overlay layer, so the delivered file is flattened rather than subtitle-preserving.
What was measured
Format support
What input and output formats, resolution limits, and duration limits are supported?
context, not decisivecapability
Input/output formats, resolution caps, and duration limits affect usability, but they are not the main quality measure for background removal or replacement. (3 of 3 judges)
What was given, what came back
Test input: Indoor Talking Head · video · group: remove-or-replace-video-backgrounds-using-ai
Input — what we sent
Indoor Talking Head
An indoor talking-head video of a person speaking directly to the camera with a static background, natural hand gestures, and facial expressions. It was used to test how well tools preserve the subject while replacing an indoor background.
Why this input is hard
- · Static camera performance
- · Face and body segmentation
- · Facial expression preservation
- · Background replacement accuracy in an indoor environment
Output — unretouched
Also checked on this input — same tool, 3 other criteria
Hair and fine detail⚠ StruggledThe hair boundary is smoothed into a flyaway-free, softened line at a third timestamp, showing the model also blurs fine hairline detail.Hair and fine detail⚠ StruggledA raised-hand gesture collapses into a soft blob with fingers and knuckles not clearly separated, showing weak preservation of fine hand detail.Temporal consistency◐ MixedThe same softened hairline reappears at a third timestamp, so the fine-detail loss is persistent across frames rather than isolated to one pose.
Provenance
- Observation
- 326b62ca-0581-4a0f-b6e0-e750a417e673
- Evidence run
- 10155228-7ce8-438a-a210-547331ef080b
- Study
- Remove or Replace Video Backgrounds Using AI
- Research task
- 86ba42c2d
- Tested at
- not recorded
- Source
- first-party
- Evidence state
- verified
- Proof shown
- input + output shown
- Cost / latency
- not captured
- Repeat run
- not captured
- Tester
- not captured
The last three rows are honest blanks, not placeholders — our capture has no field for them yet.
Query this
get_evidence({
tool: "flexclip",
scenario: "remove-or-replace-video-backgrounds-using-ai"
})MCP · mcp.aidemos.com/api/mcp
Free with attribution.
Same input, same check — 0 other tools
measured on Format support
No other tool was measured on this criterion for this input.
This evidence is published in
Real inputs and real outputs, no retouching · every cell queryable via API & MCP · aidemos.com