Quality Control
Score, scan, and gate before you commit GPU time.
<0.02s
Per frame scoring time
4
Risk flag categories
6
Metadata categories scanned
Technical overview
Preflight scoring evaluates each frame for characteristics that degrade downstream output: excessive motion blur (creates smeared depth maps), extreme occlusion (objects fully overlapping defeat depth inference), very low light (noise floor overwhelms the depth model), and unusual aspect ratios or crops that fall outside the training distribution of the depth model.
Each frame receives a fitness score and a set of risk flags. You can configure minimum thresholds — frames below the threshold are skipped or replaced with the nearest passing frame. This prevents the common failure mode where 95% of output is excellent but 5% has visible artifacts, forcing a full manual QA pass.
Metadata privacy scanning runs at the container level before frame extraction. It detects GPS coordinates, device serial numbers, camera model identifiers, creation timestamps, and software version strings embedded in EXIF, XMP, and container metadata. You can choose to strip all metadata, strip only sensitive fields, or just generate a report without modifying anything.
Use cases
- Score frames before depth estimation to avoid wasting GPU time on poor candidates
- Flag and skip motion-blurred frames that will produce smeared depth maps
- Detect and redact GPS data from personal footage before sharing or uploading
- Generate a metadata audit report for compliance review before distribution
- Set quality gates to ensure consistent output across batch runs
Configuration
- Minimum fitness scoreslider
Frames scoring below this threshold are skipped. Range: 0 (accept all) to 1 (only perfect frames).
Default: 0.3
- Risk flagsselect
Which risk categories to check: motion blur, occlusion, low light, unusual crop.
Default: All
- Metadata scan modeselect
Strip all metadata, strip sensitive only, or report-only (no modification).
Default: Strip sensitive
- GPS redactiontoggle
Specifically remove GPS coordinates from EXIF and XMP metadata.
Default: On
Pipeline stages
Detects shot boundaries, motion complexity, and optimal processing parameters before any frames are touched. This preflight pass prevents wasted compute and flags potential issues early.
Adaptive sensitivity with configurable shot boundary detection.
Output formats
Per-frame fitness scores, risk flags, and recommendations.
List of detected metadata fields with sensitivity classification.
Source file with sensitive metadata stripped. Original format preserved.
Workflows
Preflight-gated batch processing
Enable preflight scoring on a batch run. Frames that score below your threshold are skipped, so the output only contains frames that will convert cleanly. The skipped frames are logged for review.
Metadata audit before distribution
Run the metadata scanner on a folder of files before uploading to a platform or sending to a client. The report shows exactly what personal data is embedded in each file, with one-click redaction.
Quality threshold tuning
Process a sample batch at different threshold levels to find the right trade-off between coverage (how many frames pass) and quality (how clean the output is). Lower thresholds include more frames but allow more artifacts.
Industries using this
Start with a free desktop install.
Desktop processing is free with no job limits. Pro adds cloud processing, watermark removal, and advanced 3D controls.